[joshuafkon]

I thought the Financial Times had a more in-depth analysis of the structural reasons behind the spike in the REPO market than I've seen elsewhere. I've seen a few sources point out the tax payment due on the 15th, and the settlement of a large treasury sale, but as the article says:

"...Analysts say these two things alone should not cause the deep cracks in the repo market that we have seen this week. The underlying issue is more structural.

The Fed has been reducing the size of its balance sheet, letting the Treasuries and mortgage bonds it bought following the financial crisis roll off. In turn, that reduces the amount of cash reserves banks hold at the Fed... 'We have had tax payments in the past. What is different this time is that it has followed a period of quantitative tightening,' said Jon Hill, an interest rate strategist at BMO Capital Markets. 'Companies sucking cash from the market was just the tripwire that brought things falling down.'”

[beezle]

My understanding is that Treasury took in an additional 80+ in cash against their stated goal of having 350B cash on hand (by quarter end I think). Obviously this is a very large draw of liquidity out of the system and is likely the single biggest reason for the squeeze.

Not quite 2/3 of collateral at the Fed operations has been treasuries the rest mortgage paper with a tiny amount of agencies.

The financial system/world runs on repo and it is troubling this is happening at all. My gut is Mnuchkin knew this would happen and is trying to force the Fed into backdoor easing through another round of QE before gradually releasing the cash back into the system by slightly lower than expected issuance of new debt.

[nickles]

> The Fed has been reducing the size of its balance sheet, letting the Treasuries and mortgage bonds it bought following the financial crisis roll off. In turn, that reduces the amount of cash reserves banks hold at the Fed.

I'm somewhat skeptical that the reduction in Fed balance sheet is the impetus for the recent surge in repo rates, especially since the Fed ended its balance sheet unwind in August.

From August 2014 to January 2018, the Fed held $4.4 trillion in assets [0]. From 2014 to 2016, overnight repo rates remained stable and low [1]. In December 2015, the Fed began its hiking cycle. From this point onwards, repo rates began drifting up, consistent with an increasing fed funds rate [2]. It would appear that the changes in repo rates were not driven by Fed balance sheet during this period.

Starting in January 2014, banks held $2.4 trillion in excess reserves at the Fed, peaking at $2.7 trillion in August 2014 [3]. By January 2018, excess reserves held at the Fed had fallen to $2.1 trillion, independent of any change in the size of the Fed's balance sheet. In 2016, repo rates remained steady, despite a $400 billion decrease in excess reserves. In 2017, overnight repo rates drifted up, despite an increase in excess reserves of $200 billion. Here, repo rates do not exhibit any obvious impact from fluctuations in excess reserves.

As Fed balance sheet declined by $700 billion in 2018, excess reserves declined by roughly the same amount. However, even as Fed balance sheet continued shrinking in 2019, repo rates held fairly steady, corresponding to the stable level of the effective fed funds rate.

Given all this, it doesn't look like reductions in Fed balance sheet have been the sole driver of declining excess reserves, nor does it appear that the quantity of excess reserves correlates strongly with overnight repo rates. Furthermore, the Fed announced an August 2019 conclusion of the balance sheet unwind in July's FOMC statement. Finally, excess reserves held at the Fed are now 1000x greater than they were in 2007.

In short, the recent spike in repo rates happened despite a massive overhang of excess reserves and in the absence of a shrinking Fed balance sheet. I wonder if the decline of the interbank loan market, which serves as a source of short term funding for banks, is related. At the time the Fed discontinued reporting interbank loan volume (2018), volumes had declined ~75% from precrisis levels [4]. Such low volumes were last seen in 1979. This, in part, explains the large amount of excess reserves held by banks. If they do not have confidence they will be able to borrow when they need to, they must maintain such reserves.

[0] https://www.federalreserve.gov/monetarypolicy/bst_recenttren...

[1] https://tradingeconomics.com/united-states/repo-rate

[2] https://fred.stlouisfed.org/graph/?g=URW

[3] https://fred.stlouisfed.org/series/EXCSRESNS

[4] https://fred.stlouisfed.org/series/IBLACBW027NBOG

[yasp]

Any idea why the Fed discontinued data series #4 and what a good replacement for it is now?

[markus_zhang]

I'm wondering why this hasn't affected the stock market.

[dragonwriter]

It probably has; if you think it hasn't, you probably have an incorrect view of what the stock market would do without it.

[markus_zhang]

Makes sense.