The Federal Reserve’s balance sheet has ballooned from around 870 billion dollars worth of assets in 2007 to near 8 trillion dollars currently. Despite such a monumental increase, inflation and economic growth have been both dull and underwhelming during that time frame.

It seems a quandary as to why despite rounds and rounds of Quantitative Easing, fiscal stimulus, and various interventions that inflation along with economic growth have remained stubbornly low. Despite exorbitant excess reserves and even Jerome Powell admitting that he “flooded” the system with liquidity, inflation has only recently began to rise up. Even that too is marred with ambiguity as bond yields are falling once again. That alongside various other indications have been pointing to the conclusion that the long overdue inflation the Fed has been adamantly striving for may be transitory, and the recovery short-lived. There is a plethora of angles one could take in order to explain the Fed’s inefficacy, however for the intents of this paper I will focus on the Fed’s favored policy of Quantitative Easing.

Quantitative Easing, simply put, is an asset swap. Wherein assets are exchanged for bank reserves. Assets typically take the form of government bonds but can also include corporate bonds, mortgage-backed securities, and even stocks. This provides banks with liquidity which can be lent out, expanding the money supply while simultaneously lowering interest rates. This expansion of the money supply and lowering of interests rate would, ceteris paribus, be inflationary as investment is spurred and spending is increased as consumers attempt to maintain their current level of cash balances. However believing in this assumption as many central bankers do, is not seeing the forest through the trees. While much focus has been put on the level of reserves and liquidity banks have, very little effort has been allocated to the investigation of the flip side of these large scale asset purchases. There has been much less emphasis and research comparatively on what role collateral plays within the financial system.

It is imperative to understand that collateral, these securities that are extracted through the process of QE play a pertinent and integral role in money creation that underpin the entire financial system. Furthermore, theses assets are heterogenous by nature and have varying degrees of liquidity, desirability, and blank.

For example, the demand for safe, liquid assets has increased greatly over time in part due to the rise of institutional cash pools. These cash pools are centrally managed cash balances of large corporations and institutional investors, including the likes of asset managers and pension funds (Pozsar 2013). Looking at the graph below, it is clear to see the expeditious accrual of these idle dollars.

However research by Zoltan Pozsar has shown, typical M2 quantifiable money does not satisfy the pecuniary demand of the cash pools. That is because these managers do not desire money for its medium of exchange value, rather for its investment purposes. Additionally there are a plethora a legal obstructions that bar these cash pools from holding their money in typical deposits. It logically follows that typical bank reserves do not adequately satisfy money demand. Thus, facing a systemic shortage of safe liquid assets, namely government bonds, the Eurodollar system turned to securitization and collateral intermediation to cope with the excessive demand. Pozsar writes,

“At any given point in time, the volume of maturity transformation conducted outside the traditional banking system is closely related to the shortage of offshore-term, government guaranteed instruments relative to the volume of institutional cash pools.”

The figure below illustrates this observation, showing how financial innovations within the shadow banking system attempted to equilibrate with the excess demand for these near-moneys.

Jeff Snider, head of global investment research at Alhambra Investments, has done extensive research on the Eurodollar System (I strongly recommend his work and podcast with Emil Kalinowski) and comes to similar conclusions in regards to the systemic disinflationary pressures crippling the post 2008 monetary environment. In a recent article he explores these different aggregates of money to elucidate why despite so much alleged money printing and the flood of liquidity that could rival the days of Noah, inflation has yet to stick.

For reference the dark grey region of the graph represents commercial paper, which serves as a proxy for asset backed securities here (ABS). ABSs are a product of securitization and one of the shadow banking system’s methods of creating private alternatives to pristine government-backed collateral. The increase in checkable deposits is dwarfed by the vaporization of these other near-moneys. Just gazing in the shallows of the shadow banking system allows us to see the pecuniary destruction that would lead to deflationary undertow.

These findings are consistent with the body of research that attempts to quantify the size of this global shadow banking system. Various private sector estimates varies from $13 trillion to $60 trillion, however what is for certain is the dramatic decline in size after the Great Financial Crisis. After 2008, this wholesale funding network fragmented and become systemically impaired. The drying up of useful collateral (mortgage-backed securities), constrictive regulation (Dodd-Frank and various balance sheet constraints), and the perversion of the collateral allocation process and interbank trust all played contributing roles in unravelling the Eurodollar System. Thus it is hopefully clear to see 1) that money printing that the Fed has not ameliorated the monetary destruction plaguing the modern monetary environment and 2) it is increasingly non-bank institutions supplying liquidity into financial markets, these non-bank institutions cannot hold bank reserves, instead preferring various near moneys like Treasuries.

I now wish to focus on collateral intermediation for the purpose of illustrating why there is an immense demand for pristine collateral throughout the banking system, and how collateral takes on the role of the medium of exchange when bank reserves cannot.

Collateral intermediation benefits the real economy by facilitating financial transactions and lending. Given a general deficiency of collateral, or even wrong types of collateral, intertemporal coordination becomes distorted as a suboptimal allocation of wholesale funding and consumer savings occurs. This suboptimal allocation may very well result in a general decline in interbank activity and lending as banks cannot put up the sufficient collateral to enable the desired transaction. This accompanied with various other factors help explain the lack of lending and inflationary pressure after 2008.

One of the primary methods of obtaining funding with the shadows of the banking system, or colloquially the Eurodollar System, is through repo agreements. These banks and various financial institutions find funding for their daily operations by posting their collateral to obtain a securitized loan. Typically, collateral rich institutions like pension and investment funds obtain collateralized loans through a money dealer, which usually take the form of a wholesale bank. While on the flip side these institutional cash pools will provide the liquidity to these dealers as they prefer to keep their wealth in the form of safe, liquid assets. These wholesale banks play a vital role in the optimal and efficient allocation of collateral and liquidity dispersion, and its clear to see how any apprehension from these money dealers could impinge the entire monetary system. Jeff Snider, once again, illustrates succinctly what I’m outlining. However, it is important to note that this outline is extremely generalized as the real system is much more intricate and complex. In the figure below, Snider maps out how collateral flows from these pools into the hands of money dealers, wherein collateral can be transformed, rehypothecated, and allocated efficiently. A crucial point here is that money dealers are not completing transactions and balancing their balance sheets on a 1 to 1 basis, rather they are transacting an array of assets against an array of liabilities that needn’t cancel each other immediately.

A critical distinction of the post GFC1 environment is that central banks now also possess silos of collateral. This collateral lay mostly idle in the Federal Reserve’s SOMA portfolio, thus also hampering the efficient allocation of collateral to finance funding and lending.

Now we are beginning to see the forest through the trees, yet there is still much more to unravel. One of the more intriguing elements of this collateral intermediation is the rehypothecation of collateral. While the briefly mentioned securitization is also pertinent to this discussion, its use has largely fallen off compared to collateral based operations after 2008. Where the same piece of collateral, whether it be a treasury bill or a junk bond, is repledged and reused to finance multiple collateralized loans.

I’ll use a simple thought experiment to illustrate this dynamic.

Say a hedge fund desires funding to cover a previous debt or position, so it goes into the repo market and finds a money dealer. The money dealer gives the hedge fund the funding it desires on an overnight loan. This loan is collateralized and allowed to be reused in order to get the cheapest rate possible, meaning the hedge fund gives the money dealer an equivalent amount of collateral (plus a haircut) as “insurance” for the loan. The following day the hedge fund will come back and pay back the loan and take back their collateral, before that however, the money dealer will reuse, or rehypothecate this collateral to another party to obtain funding for a variety of financial operations. So lets say that hedge fund pledged a US Treasury bill, the money dealer takes that T-bill and repledges it to another financial institution. Then that financial institution will repledge that T-bill again to fund their daily operations and so on and so forth.

We can see how there is a “collateral multiplier” in the same vein as the textbook money multiplier effect. This collateral velocity underpins much of the financial system and has real economic effects. Furthermore, the collateral itself is taking on the medium of exchange role within the wholesale interbank lending markets. There are various examples of banks, even whole markets zealously overbidding for treasuries, indicating the intrinsic value they possess in this modern ecosystem.

Quantifying the collateral multiplier allows us to gain invaluable insight into how the shadow banking system is performing. Some research has been done to uncover this collateral multiplier, but this is an incredibly difficult proposition as there is such little data on this. Manmohan Singh’s research found that the total volume of collateral transactions and general volume of collateral both fell substantially after 2008, which is consistent with our previous observations. More modern research places the multiplier between 6 and 8. In the figure below illustrates the collateral multiplier for U.S Treasuries, the most pristine collateral.

Furthermore, independent research of dealer banks confirm these findings. One intriguing line of research is the discovery that local maximums of the collateral multiplier correspond to what Jeff Snider refers to as Eurodollar crises, which are periodic dollar shortages which plague the post 2008 environment. (For further information on these events I once again recommend his work)

What matters here is that Snider’s claim holds under empirical scrutiny, that much of these reflationary sub cycles where reuse increases while total collateral is largely unaltered have to deal with dealer bank’s perceptions of risk and fragility. The next logical question is why, despite this ostensibly healthy collateral chain, excess reserves, and a competent central bank would this system be marred with fragility?

While this collateral intermediation and it’s multiplier effect are invaluable to wholesale funding, it is not without its respective pernicious effects. Going back to our example, lets say perceptions of risk change, (or useful collateral dry’s up like in 2008) . Our hedge fund now needs funding to pay off the debt it accrued from the financial institution. So the hedge fund goes to another money dealer it is acquainted with in attempt to find funding. However due to the increased perception of risk, the dealer bank requires a qualitative increase (a shift from for example a corporate bond to a treasury) or a quantitative increase (simply more collateral) in order for the money dealer to deem the transaction worthwhile given their balance sheets considerations. The hedge fund is not able to cope with this change and is forced to liquidate their assets to pay off their debt or default. This drives the dollar up, exacerbating the issue. While simultaneously, margin calls typically ensue as the financial institution will look to sell the collateral immediately. In the short term the price of the collateral will fall, which may very well lead to more margin calls and more seizure of collateral. This spiraling effect is amplified by the length of the collateral chain, or how many times that collateral had been repledged and reused. Additionally, the hedge fund needn’t have even owned that piece of collateral it lent out, it may have obtained that collateral from a reverse repo agreement. The distinction of ownership becomes very blurred here as increased collateral chains increase the interconnectedness of the financial system, leading to systemic risk of collateral runs and inherent fragility. We can view these perception tendencies through the figure above. As perceptions of risk increase, the volume of lending and repledging decreases, drying up the financial markets and manifesting deflationary effects.

It seems the system is in a catch-22, where by decreasing the interconnectedness and fragility (decreasing velocity) the systems liquidity dries up and leaves wholesale transactions frustrated. This has delirious effects on the real economy (that are admittedly hard to pinpoint)

Now lets bring all of this back to the Fed and QE. How do these large scale asset purchases that the Fed uses affect this wholesale system? When the Fed strips treasuries and other assets from the system, it is stripping away the valuable collateral that underpins the entire system. Furthermore we can logically deduce that banks would be forced to increase the already existing collateral chains to maintain the same level of activity, increasing the fragility and interconnectedness of the system. Recent research from the Federal Reserve itself have begun to analyze this process, a 2020 paper found

“A $33 billion dollar increase in weekly Fed purchases leads to a 0.5 increase in the collateral chain”

Given the Fed’s activity in 2020, that would suggest a 0.66 increase in the collateral multiplier. So the remaining collateral in the system is stretched out 0.66 more times on average just to maintain status quo.

Viewing this from another angle, Jeff Snider has taken the time to analyze data from the Federal Reserve bank of New York on triparty repos. He finds that during reflationary periods, a wider variety of collateral is put up in these repo agreements. While conversely during these Eurodollar crises, or periodic dollar shortages, the opposite generally ensues. This is what we would expect, he states, as economic conditions falter the arrays of collateral being used in repos should condense into safer, more liquid assets.

On the graph above it is clear to see the number of observations (variety of collateral) increases during these reflationary periods, while during Euro$ events they stagnate or fall. Furthermore he points out that the overall increasing trend of observations suggests that the Federal Reserves persistent removal of useful collateral like treasuries and MBSs forces banks to use a wider variety (and more quantitatively) of collateral to maintain their operations.
Taking these remarks to a logical conclusion, the aforementioned Federal Reserve paper sums it up nicely,

These observations combined suggest that the central bank can effectively reduce the interconnectedness of the financial system by reducing the size of its balance sheet.

By selling treasuries, not buying them, can the Fed help extirpate the inherent fragility within the system. This seems counterintuitive, as the primary Keynesian solution for downturns in money injections like QE, and furthermore even in a Monetary Disequilibrium lens an excess demand of money should be offset by an increasing supply of money. Of course that is exactly what is being prescribed. As I’ve mentioned before, money in this topic is heterogenous, in a similar way that Austrians describe capital as heterogenous. So just as the Austrian Business Cycle Theory accentuates mal-investment versus over-investment, the distinction between different types of money is of vital importance. Haphazard money printing measured in aggregates is futile in this regard and simply profligacy. Instead, focusing on the types of money that will satisfy demand is a much more effective tactic. By unwinding the Fed’s balance sheet, and releasing Treasuries into the market, the most pristine form of collateral, the Fed would be directly mitigating the excess demand and fragility of the system brought on by overstrung collateral chains. This solution isn’t of course without unideal side effects, so further analysis of this proposition is warranted.

Furthermore, research has shown that central banks typically have a bias towards illiquid and precarious forms of collateral, as they typically buy up all the unwanted forms. This induces a positive feedback loop as it incentivizes the production of those collateral types, inducing suboptimal intertemporal allocation of savings which will exacerbate business cycles. We can additionally apply Gresham’s Law to collateral in order to further illustrate this concept. Just as Gresham’s ‘good money’ will be used deliberately or trade at a premium, pristine collateral (US Treasuries) will act the same. Banks will save their most pristine collateral for bilateral repos, intermediate collateral for tri-party repos, and their worst collateral for transactions with their central bank. As banks use the pristine collateral sparingly, the increased usage of the ‘bad collateral’ is something we can empirically observe.

In the blue are the pristine forms of collateral, while the green colors indicate the ‘bad’ forms of collateral. We can see how despite the ‘good’ collateral making up a large section of usable assets, only a fraction of that is actually used in collateral based operations. Meanwhile the ‘bad’ collateral is used quite a bit more. This is simply one more angle to view the distinct problem this system has.

An immediate response from a proponent of QE would be that is accomplishes its goal, that it lowers rates to increase lending. So by doing the opposite we would stifle recovery. To the contrary, given what I’ve laid out I would argue that it is the stripping away of collateral and increasing of collateral chains that exacerbates the demand and lowers yields. Not necessarily the Fed’s direct buying. Even New Zealand’s central bank came out and stated they found that a 10% of GDP QE program lowered rates by 50 basis points. This extremely pitiful number lends credence to the claim that it is the negative side effects of QE that lowers yields, not the “stimulating” direct ones. The financial sector is already flocking towards treasuries and other safe, liquid assets, the FED, merely adds insult to injury.

In short, Quantitative Easing has extremely detrimental effects that distort the underpinnings of the entire financial system. That is because bank reserves are only half the story. While dollars serve as a medium of exchange for traditional banking and retail purposes, it is this ever so disregarded collateral that facilitates exchange in the shadows. However, while unwinding the Fed’s balance sheet and supplying more treasuries should help to alleviate short term stress, it does not solve everything, nor does it come close to doing so. Much of the problems stem from the system itself and its fracturing during the Great Financial Crisis. However that is a discussion for another time.

This modern system can be likened to the beach, with periodic reflationary waves surging towards the shoreline, while strong deflationary riptide pressures pull the waves back into the ocean. As we all float upon the sea of liquidity aboard Powell’s Ark, the inflationary waves draw all the attention and headlines. Yet it is the deflationary riptide that warrants concern. The longer this murky undertow goes unnoticed, the further we will all be drawn out to sea.