Is the Liquidity Problem in the Repo Market Portentous of a Coming Market Correction?
Or was it merely a technical blip?
Déjà vu
For those of us who have not only lived through the last two massive market corrections but worked within, or somehow touched, the banking system, several events are starting to give us a sense of déjà vu of “peak market” and a potential downward spiral.
Perhaps the most obvious was the WeWork débâcle. The blind business plan optimism, massive over-valuation and the fact that professional investors were blindsided is reminiscent of the behaviour of both companies and capital providers at the peak of the Dot-Com bubble in 2002.
More menacing is the liquidity problem we witnessed in the repo markets in September 2019. But was that merely a blip or a sign of trouble ahead?
Not the repo market again!
The US Federal Reserve has been injecting billions of dollars of cash into the repo market since September 17th¹.
The repo market is where banks and a select group of institutional investors (more on this later) go to obtain short term capital. Typically banks with excess cash lend at very short durations (overnight to weeks) to other banks who put up Treasury securities (such as T-bills²) as collateral. The repo market was also traditionally where the Federal Reserve conducted its open market operations to ensure that the short term interest rates (like the overnight repo rate) matches the target rate that it has set.
The repo market typically moves with a daily range of 10–20 basis points (bps; 100bps = 1%), with a daily turnover of about $1 trillion. On September 16th through to 17th the overnight repo rate spiked about 700 bps (!), from around 2.25% to over 8.5%. (At the same time the Fed Funds rate increased by 5 bps to 2.30% which was above the Fed’s target, the first time this has happened since the 2008 financial crisis.)
This spike in the overnight repo interest rate means that there was an unexpected shortage of liquidity in the financial system³. Essentially the banks in the repo market were reluctant to lend to each other, hoarding cash instead. This was reminiscent of the 2008 financial crisis where one of the early signals (as early as August 2007⁴) of a problem with the banking system was that banks had refused to lend to each other, resulting in a sudden spike in the overnight repo and LIBOR rates.¹⁴
On the morning of Tuesday September 17th, the New York Fed announced an injection of $50bn to stabilize the repo rates⁵. This was the first significant cash injection by the Fed since the 2008 crisis. On September 18th and 19th an additional $75bn was injected into the markets. The next day, on September 20th, the Fed announced a series of overnight and short term repos through October 10th. Then on October 30th the Fed committed to “keeping bank reserves at appropriate levels” (that is ensuring liquidity) through both temporary open market operations and by purchasing T-bills at least through January 2020⁶.
That is, by October, the Fed had increased the overnight repo limit to $120bn from $75bn, lent for longer periods on the repo market to the tune of $45bn and started a “permanent” operation to buy short-dated Treasuries at $60bn/month.⁷
While the overnight repo and US LIBOR rates have fallen back down to pre-September 16th levels, the fact that such a large amount of liquidity had to be continuously infused into the banking system should make us wonder a) why was there an unexpected shortage in the supply of liquidity; b) why was there a reluctance to lend and; c) was there unexpected demand for cash?
Unexpected shortage in the supply of liquidity — a result of structural changes since the 2008 crisis?
One of the structural changes that happened during the last financial crisis was the intended and unintended effects of Quantitative Easing (QE). During QE the Fed bought long-dated Treasuries from banks, resulting in banks becoming awash with cash to the tune of nearly $2.5 trillion at the peak in 2014.⁸
In addition, prior to 2008, the Federal Reserve has been debating and instituting changes to reduce the competitive disadvantage of depository institutions relative to other financial institutions. Traditionally the legal reserves banks held at the Federal Reserve Bank attracted zero interest; in 2008 the influx of cash due to QE resulted in tremendous pressure on the fed funds rate. To solve the problem, in October 2008, the Federal Reserve began paying interest on reserves (known as the Interest on Excess Reserves, IOER) and operating on a “corridor framework” ⁹.
This fundamentally changed the liquid-asset holdings of US banks. One unintended consequence according to the Economist newspaper was that the Fed lost the ability to gauge how much liquidity the banking system needed on a daily basis. Since the banks have been awash with cash, and excess cash held at the Federal Reserve Bank attracted an interest, the repo market operations by the FOMC essentially became redundant¹⁰. It was through the repo market that the Federal Reserve gauged how much liquidity the banking system required.
This blindness came at a time when the Federal Reserve was reducing its balance sheet, which at its peak in 2017 stood at $4.7 trillion (pre-2008 QE the balance sheet was under $1 trillion). One potential explanation for the shortage of supply was that the Federal Reserve underestimated demand and overshot in contracting its balance sheet.
Was the reluctance to lend a result of structural changes since the 2008 crisis?
Quantitative Easing, the contraction of the Federal Reserve’s balance sheet from October 2017, and the Federal Reserve’s interest on excessive reserves have resulted also in a concentration of liquidity in the 4 largest banks in the US banking system: Bank of America, Citigroup, JPMorgan Chase and Wells Fargo. These 4 banks in essence have become the provider of funds to the repo market (their net lending positions hitting $300bn at the end of June 2019)⁸.
As these banks started amassing Treasury securities both through the Fed’s balance sheet contraction and new issuance (we will look at the demand side below), they accounted for more than 50% of all US Treasuries held by US banks according to the Bank for International Settlements (BIS) at the end of the second quarter in 2019. However, at this point they held only about 25% of the supply of liquidity in the repo market.
It is possible that these banks had decided to stop lending as their reserves dwindled (while they accumulated Treasuries) at a time when some pressure was applied to the repo market for liquidity. Since the net lenders were only 4 institutions, it could have been far easier for liquidity to dry up, in contrast to a system where lending and borrowing was more heterogeneously distributed throughout all the participants.
One other often cited reason for the reluctance to lend is that banks have been conditioned by regulation to be overly cautious following the 2008 financial crisis. This must have been exacerbated since the Fed began paying IOER, since relatively speaking it is now easier to justify hanging on to those reserves rather than use them in the repo market when there is a whiff of increased risk or volatility and regulatory requirement thresholds are near. If this were the case, again having the liquidity of the system concentrated in a few institutions would heighten conservatism.
However, as important to the supply of liquidity issues is the demand for liquidity, and here is where active dangers may lurk.
What caused the unexpected level of demand for liquidity?
It is often repeated in the popular press that the overnight repo market tightened on September 16th because of a confluence of a couple of mundane factor: that corporations required cash for tax payments and that cash was required for the settlement of purchases of Treasury auctions. However, the Federal Reserve has been dealing with volatility created by these 2 events without drama for decades.¹¹ So what is going on?
The US’s fiscal deficit will reach $1 trillion next year (or 4.7% of GDP). The Treasury is funding this by borrowing more through the issuance of Treasury securities. However, foreign buyers are shying away and buying proportionally less of the auctioned Treasury securities.
It is possible that one factor was the demand for cash to settle Treasury purchases, but there is probably a twist.
The Economist newspaper estimates that foreigners now hold 35% of Treasuries compared to 40% in 2017.¹² This slack is taken up by the Primary dealers¹³ who now hold $250bn of Treasuries and finance an additional $2trn through repo.¹⁰ It is likely that demand from the Primary dealers could not be matched with supply on September 16th causing the spike in rates.
In essence, this was a secular trend (increased borrowing by the US government, decreasing foreign participation and increasing Treasuries purchases by Primary dealers) colliding with the reduction in cash reserves in the banks described earlier in the article. But why was this not foreseen?
Because that’s likely not the full story…
One other change since the last financial crisis was that the Fed added investment funds and non-bank securities dealers to the list of approved borrowers in the repo market (see footnote 10). In 2014 non-bank investors including hedge funds and real estate trusts were already the biggest users of the repo market according to reporting by the Financial Times.¹⁵ The Bank for International Settlements theorised that the September 16th liquidity problem was due to demand for cash from these highly leveraged institutions.¹⁶ Hedge funds in particular are using the repo market to fund their search for yield, particularly in derivatives.
Which brings us almost full circle to one of the major causes of the last financial crisis — derivatives.
A blip? Unlikely.
The low interest rate environment, Quantitative Easing and its unwinding, changes to the plumbing of the repo market in the US, increased borrowing by the US Government while foreigners have reduced their appetite for US Treasuries, increased participation by highly leveraged institutions in the repo market (potentially increasing counterparty risk to cash-rich institutions) have likely all contributed to the environment which led to the seizing up of liquidity we witnessed in September.
What it tells us is that we now have a fragile banking system and are probably starting to face problems at the other end of the yield spectrum where derivatives such as mortgage, student loan, automobile and other securitised instruments live.
The “corridor framework” combined with low interest rates hamstring the Fed to fight any fires of reasonable scale. In order to inject more liquidity in the system, another significant asset purchase programme could be started. However this would serve not only to increase the Fed’s balance sheet again, but also to pump up the asset-valuation bubble in which we live in now even more.
If the plumbing cannot be easily or quickly repaired, the risks of a crash does seem to be higher, especially in the current environment. Indeed, the sudden lack of liquidity in the repo market on September 16th & 17th, followed by intervention by the Fed, might someday be seen as the first sign of the next big correction, even if that correction is still many months away.
End Notes
¹ Fed announces plans to provide more support for the repo market. AP News, 20 September 2019 (https://apnews.com/efb5eaaa848c43d081fc91f136909e95).
² T-bills are short-dated Treasury securities issued by the US Government.
³ The keyword here in “unexpected” which we will explore further in this article.
⁴ On August 9, 2007, the ECB had injected €94.8bn ($130bn) into the banking system when faced with a similar tightening of the short-term credit market. This amount, to be followed later by more cash, was more than the ECB had injected after 9/11!
⁵ Note that the overnight repo rate had actually started to spike on Monday September 16th.
⁶ Federal Reserve Implementation Note issued October 30, 2019. (https://www.federalreserve.gov/newsevents/pressreleases/monetary20191030a1.htm).
⁷ See also https://www.cnbc.com/2019/10/23/fed-repo-overnight-operations-level-to-increase-to-120-billion.html
⁸ September stress in dollar repo markets: passing or structural? BIS Quarterly, 8 December 2019 (https://www.bis.org/publ/qtrpdf/r_qt1912v.htm)
⁹ Monetary Policy Under a Corridor Framework, Kansas City Fed Economic Review, 4th Quarter 2010 (https://www.kansascityfed.org/publicat/econrev/pdf/10q4Kahn.pdf)
¹⁰ Why the repo market went awry…and how the Fed should fix it. The Economist, 2 November 2019 (https://www.economist.com/finance-and-economics/2019/11/02/why-the-repo-market-went-awry)
¹¹ See Ben Bernanke’s 2005 speech on Implementing Monetary Policy where he actually mentions these same 2 events that cause exceptionally large payment flows. (https://www.federalreserve.gov/boarddocs/speeches/2005/20050330/default.htm)
¹² See footnote 10. Also see the Congressional Research Service’s Foreign Holdings of Federal Debt updated July 26, 2019 (https://fas.org/sgp/crs/misc/RS22331.pdf)
¹³ Primary dealers are banks approved to deal directly with the government during Treasury auctions.
¹⁴ See: Deciphering the Liquidity and Credit Crunch 2007–08. Brunnermeier, National Bureau of Economic Research, Working Paper 14612. (https://www.nber.org/papers/w14612.pdf)
¹⁵ See also: Biggest investors replace banks in $4.2tn repo market. Financial Times, May 29, 2014. (https://www.ft.com/content/ca529c5e-e5db-11e3-aeef-00144feabdc0?siteedition=intl#axzz33CXfVXjP]
¹⁶ See also footnote 9 where the BIS has market intelligence on this issue.
