Repo dislocations – a case study in market inefficiency

A little known but important segment of fixed income markets – known as the ‘repo market’ – has received an unusual amount of attention since the latter part of 2019.

The repo market is where financial institutions borrow and lend cash for short periods to manage their funding needs. In the US alone, the average daily repo amount outstanding is c. $3.9 trillion and daily turnover amounts to c. $1 trillion.

As the Bank for International Settlements (BIS) explains:

“A repo transaction is a short-term (usually overnight) collateralised loan, in which the borrower (of cash) sells a security (typically government bonds as collateral) to the lender, with a commitment to buy it back later at the same price plus interest.

Repo markets redistribute liquidity between financial institutions: not only banks (as is the case with the federal funds market), but also insurance companies, asset managers, money market funds and other institutional investors.

In so doing, they help other financial markets to function smoothly. Thus, any sustained disruption in this market, with daily turnover in the US market of about $1 trillion, could quickly ripple through the financial system.

The freezing-up of repo markets in late 2008 was one of the most damaging aspects of the Great Financial Crisis (GFC).”

– BIS quarterly review, Dec 2019

The repo market is an important subset of the broader short-term money markets, which are effectively the plumbing of the global financial system. Trillions of dollars flow through money markets daily as a diverse array of market participants borrow and lend cash for short periods, effectively setting the price of cash in each currency.

If markets were efficient, the pricing of cash would be consistent across all the various instruments that trade in money markets, including repo rates. However, recent events in the US repo market are a stark reminder that interest rate markets are in fact not efficient and if anything, are becoming more inefficient.

While this growing inefficiency causes problems for some, it creates opportunities for others. (details here)

What went wrong?

Repo markets normally don’t get attention outside a small club of specialists because they generally function smoothly. To the extent there is volatility, it’s usually not large enough to warrant broader scrutiny.

Things changed in September last year when the overnight US repo rate shot up as high as 10%. This meant large financial institutions were paying an interest rate of 10% p.a. to borrow short term USD cash, even though other closely related short-term USD interest rates, which are close substitutes for repo, remained in the 1.5 – 2.5% range.

This is all the more unusual because the overnight repo rate is essentially a risk-free interest rate, which makes 10% an absurdly high level.

Some short-term volatility in repo rates around quarter and year ends is normal. It’s caused by temporary demand-supply imbalances, for example as banks get their balance sheets ready for reporting to regulators. However, this volatility has been increasing since late 2018, culminating in the unusually large September 2019 spike.

Why did it happen?

The underlying dynamics that led to the September repo spike had been building since the US central bank (the Federal Reserve – FED) began to scale down its quantitative easing (QE) program back in late 2017.

The starting point to understand all this is the concept of bank reserves, a simplified explanation of which follows.

In the same way that we all hold a certain amount of cash on deposit in our bank accounts to facilitate payments, banks have to do the same. In the case of banks, the deposits are held with their national central bank (e.g. the FED in the US) and are known as bank reserves. As part of their regulatory framework, many banking systems require that a minimum amount of reserves always be held with the central bank. This is known as a ‘minimum reserve requirement’ and is calculated based on the amount of loans a bank makes (i.e. more loans means more reserves are required). Any additional amounts held above the minimum are known as ‘excess reserves’.

The purpose of minimum reserve requirements is to ensure that payments between banks can be made reliably and that money markets, which are the plumbing of the global financial system, can function smoothly. Without this system, banks could find themselves short of cash to settle payments with each other, which in turn would cause payments in the broader economy to fail (e.g. for wages, payments to suppliers etc.), leading to financial panic, recession etc.

Additionally, in many economies reserve requirements are part of the monetary policy toolkit that allows central banks to influence the level of interest rates. For example, when central banks reduce reserve requirements banks don’t need to hold as many reserves, which means those funds are freed up to lend to banks’ customers. This increases the supply of funding available in the economy, which then can exert downward pressure on interest rates. The opposite happens when reserve requirements increase.

A big effect of the QE program that was implemented in response to the 2008 financial crisis was to increase the amount of excess reserves in banking systems. QE involves central banks purchasing government bonds and other securities from banks. To pay for these purchases the central bank creates reserves, which are credited to the banks’ reserve accounts, thereby increasing the amount of excess reserves that banks hold. In the US for example, the level of excess reserves increased from close to zero in 2007 to a peak of $2.7 trillion in 2014.

It used to be the case that reserve requirements were the main constraint on the size of bank balance sheets. If banks wanted to increase their balance sheets by lending more, they were forced to hold more reserves to support that lending.

However, a raft of new banking regulations that have been introduced in response to the 2008 financial crisis have placed a whole range of new and complex constraints on the size and composition of bank balance sheets.

This means bank reserves are no longer just used to support lending activity but are also being used to fulfil a whole range of new regulations, such as liquidity rules. The side effect is that excess reserves are no longer truly ‘excess’ because they are now required to meet all these new regulatory constraints, which in turn means banks are highly incentivised to hold much higher levels of ‘excess’ reserves than they have in the past.

This is the key point in explaining why the September repo spike happened.

In the past, when demand for funding increased in some part of the money market, such as repo, banks with excess reserves would quickly increase their lending to take advantage of the higher interest rates that came with the increased demand. These banks were effectively using their excess reserves to increase the supply of funding available in that part of the market, which had the effect of smoothing out demand-supply imbalances across money markets, pushing down those elevated rates and generally dampening volatility.

Today, even though excess reserve balances look high, those reserves can no longer be freely deployed to take advantage of funding demand spikes because they are locked up in an alternative use, which is to comply with the complex web of new regulations that have been imposed on banks.

Just one example of these new rules is the requirement for banks to hold a minimum amount of high-quality liquid assets (HQLA) on their balance sheets. They could comply with this rule by holding eligible government bonds, however reserves are a more capital efficient way of complying and therefore a portion of those ‘excess’ reserves are actually tied up meeting HQLA requirements.

So when demand for repo funding increased in September, even though the system had enough reserves, these frictions prevented them from efficiently moving to smooth out the demand-supply imbalance, resulting in the overnight repo rate jumping as high as 10%.

Therefore the problem – and the root cause of the unusually high repo rate volatility – was not the amount of reserves in the system, but rather the lack of mobility of those reserves. They can no longer move freely to smooth out demand-supply imbalances.

JP Morgan’s CEO, explained it this way:

“JPMorgan Chase & Co. had the cash and willingness to calm short-term funding markets when they went haywire in mid-September, but the banking giant said regulations held it back.

The firm has what Chief Executive Officer Jamie Dimon on Tuesday called a checking account at the Federal Reserve. When rates on repurchase agreements spiked to around 10% a month ago – roughly four times more than what JPMorgan earns at the Fed – the bank could’ve profited by shifting the money into repo.

It didn’t. The bank, Dimon told analysts following JPMorgan’s third-quarter earnings release, needed to keep that money put so it could fulfil its liquidity requirements mandated by regulators.

“We could not redeploy it into the repo market. We’d have been happy to do it,” Dimon said Tuesday. “It’s up to the regulators to decide if they want to recalibrate the kind of liquidity they expect us to keep in that account.”

– Bloomberg News, “Dimon says regulation limited JPMorgan from calming repo market”, Oct 2019

This dynamic has been further exacerbated by the FED’s scaling back of QE since late 2017 because as they stopped buying bonds from banks, they stopped crediting banks with more excess reserves. At the same time, the US government was issuing more bonds to fund spending, resulting in banks being forced to hold more of these bonds on their own balance sheets. Banks, being highly leveraged entities, had to finance these bond purchases in the repo market, thus further increasing the demand-supply imbalance.

Could it get worse?

Heading into the 2019 year-end some feared that US repo market volatility could spill over into other parts of financial markets. The more alarmist commentaries cited the risk of a systemic panic but as it turned out, the year-end passed uneventfully.

The immediate concern was that leveraged investors (e.g. hedge funds, banks), who are reliant on repo markets to fund their bond holdings, could become forced sellers of bonds, causing disorderly price volatility in bond markets.

The other concern was around contagion, given that overnight repo rates are essentially a risk-free rate that’s closely related to other risk-free interest rate proxies such as the official US interest rate set by the FED – known as the target Fed Funds rate.

The concern was that if repo rates diverged materially from the target Fed Funds rate for an extended period, it could cause a large dislocation between the market’s pricing of short-term interest rates (such as the effective Fed Funds rate) and the FED’s official target rate, thereby compromising the effectiveness of monetary policy. Essentially, the FED would lose control of interest rates.

These concerns didn’t play out because the FED stepped in to inject sufficient cash into money markets to push repo rates back down and subsequently stated their intention to keep doing so as needed.

This is a simple solution in theory but harder to implement in practice because it is difficult to precisely gauge how much cash to inject, how quickly to do so and when to stop. A more lasting solution would be to relax some of the regulations constraining bank balance sheets. While banks would love this, there is significant opposition from policymakers.

For these reasons, while the systemic contagion doomsday scenarios are unlikely, more frequent and pronounced bouts of short-term volatility in repo markets are likely.

A problem for some, opportunity for others

Repo market volatility can cause problems for investors who rely on repo to finance their bond holdings.

A common strategy used by hedge funds and other leveraged investors is to use cash borrowed in the repo market to finance bond purchases. This repo financing is typically short term in nature, which means it has to be continually rolled over in order to keep funding those bond investments.

When repo rates become volatile, as they did in September, it becomes harder to maintain access to that repo financing, which means leveraged investors may then be forced to sell their bond investments at a loss. This is particularly problematic if those bond investments are not very liquid (e.g. corporate bonds).

At the same time, repo market volatility can create opportunities for investors who do not use repo financing for leverage.

As repo rates reflect very similar underlying risk characteristics to other closely related money market interest rates, when demand-supply imbalances cause them to temporarily diverge from other rates, the resulting pricing inconsistencies create precisely the kinds of low risk opportunities that relative value investors specialise in.

Prior to 2008, when temporary demand-supply imbalances caused pricing inconsistencies across interest rate markets, enormous bank balance sheets were quick to step in and intermediate the imbalance, thereby smoothing out the pricing inconsistency and generating profits for themselves in the process.

For example, if they saw repo rates start moving higher, they would quickly respond by reallocating their balance sheets from other short-term cash instruments to profit from higher repo rates and in doing so, would smooth out the demand-supply imbalance well before repo rates reached anywhere near the 10% levels of September.

However, in the post-2008 era of highly regulated and constrained bank balance sheets, they can’t intermediate these demand-supply imbalances in the same way anymore.

Barclays explains it this way:

“… it was a surprise that banks did not step in more aggressively to provide funds than they did, even as rates spiked to very attractive levels.

We believe regulation and supervision has played a significant role in elevating banks’ preference for cash. The list of capital and liquidity requirements to which they must manage has expanded significantly since the financial crisis. “

“Even absent funding imbalances, post-crisis regulation has increased the cost of intermediation. This in turn has decreased the availability of dealer balance sheet relative to the growth of capital market and leaves the repo market vulnerable to frequent disruption.

… in our view, changes in bank regulations since the GFC have hampered the ability of banks to intermediate. The leverage ratio creates a capital charge on repo market intermediation, as this activity expands the bank’s balance sheet, thus reducing the attractiveness of this low-risk, low-return type of activity.”

– Barclays Capital, “Proposals to address repo market fragility”, Nov 2019

Recent repo market volatility is just one example of how bank balance sheet constraints contribute to inefficiencies across global interest rate markets. There are many more.

Constraints on banks mean there are now more frictions across interest rate markets that prevent capital from easily flowing to smooth out temporary demand-supply imbalances. This in turn means more pricing inefficiencies available for those with the specialised expertise and experience to profit from. (details here)