Fixing Repo: A Follow Up

Fed, Bank of Canada, Federal Reserve, Repos

In a post published here in mid-November, I traced the Fed's repo-market troubles to post-2008 changes in the importance and volatility of two of the Fed's non-reserve liabilities: TGA balances and the foreign repo pool. Then, in a companion piece, I made some suggestions for reining-in those liabilities, as an alternative to having the Fed continue to increase the size of its balance sheet—one that could actually allow that balance sheet to shrink further. The proposed reforms would even make it relatively easy for the Fed to switch from its present floor operating system to a corridor system, with a correspondingly slim balance sheet.

Since I wrote those posts, I've come across further, relevant information. Besides reinforcing my confidence in the proposals' merits, the new information also suggests that implementing them will be even easier than I'd previously supposed.

Direct Treasury Repos

In the second of my November essays, I suggested two reforms aimed at reducing the Treasury's TGA balance. Congress, I wrote,

should (1) revive the Treasury's direct repo borrowing program, expanded (as per the GAO's September 2007 recommendations) to allow all primary dealers to take part and also to allow the Treasury to offer any of its securities as collateral, and (2) restore the Treasury's "direct draw authority," or grant it permission to make use of the Fed's Standing Repo Facility.

When I wrote this, I knew only of a pilot program, begun in March 2006 and lasting just one year, allowing the Treasury to experiment with lending on the private repo-market. I noted that, though it was judged a success, the program "was extremely limited in scope and scale, with the Treasury's average of $4 billion repos a day comprising but a miniscule share of what was then a $1.8 trillion per day repo-market." Reviving and expanding the program, while making it permanent, would, I said, give the Treasury an alternative to holding large TGA balances that would yield it a higher return while being almost as safe.

Unbeknownst to me, Congress had already taken the step I recommended in 2008, by passing the "Fostering Connections to Success and Increasing Adoptions Act." What have adoptions got to do with Treasury repos, you ask? I haven't the foggiest idea. Nevertheless, it was by means of section 502 of that act that Congress chose to amend section 323 of Title 31 of the United States Code.* The revised code allows the Secretary of the Treasury to invest the Treasury's cash in

‘‘(1) obligations of depositories maintaining Treasury tax
and loan accounts secured by pledged collateral acceptable to
the Secretary;
‘‘(2) obligations of the United States Government; and
‘‘(3) repurchase agreements with parties acceptable to the

The 2008 amendment further stipulates that, instead of being obliged to place funds in any particular account, the Secretary "shall consider the prevailing market in prescribing rates of interest for investments" allowed for by the amendment.

In short, the Treasury is already free to lend all it likes in the repo-market, instead of placing money in either the TGA or in still-less remunerative TT&L accounts.

So why hasn't the Treasury ever taken advantage of this right? The answer resides in the timing of the measure granting it. That measure took effect on October 7th, 2008—one day after the Fed announced that it would begin paying interest on bank reserves. Because placing funds in the TGA reduces bank reserves, while direct Treasury repos would leave those reserves unchanged, once the Fed started paying interest on reserves, it was only worthwhile for the Treasury to prefer repos to TGA balances when repos paid less than the IOER rate. Otherwise, the Treasury would lose more through IOER payments than it could earn by supplying funds to the repo market. As the following chart shows, that was generally the case between 2008 and 2017.

Since 2017, however, repo rates have gradually risen relative to the IOER rate; and lately, they've been persistently higher—high enough to make direct Treasury repos more profitable than a large  TGA balance.

So what's been keeping the Treasury from making use of its direct repo option more recently? The main reason is the Treasury's 2015 decision to keep at least a week's worth of disbursements, and never less than $150 billion, in the TGA, as insurance against a major disruption of financial markets. $150 billion: that's thirty times the Treasury's pre-2008 minimum TGA balance! That seems like a case of "once bitten, twice shy," with a vengeance. Only the Treasury has never actually found itself unable to tap the private money market for more than a day or so—not in September 2008, or in October 2012 (Superstorm Sandy), or even in September 2011.

The Treasury claims nevertheless that it has to have such a large balance as protection against the possibility of a successful cyber-attack, or other major disaster. But my other proposed reform—reviving the Treasury's  "direct draw" authority so that it can borrow from the Fed during crises— would also address the Treasury's concerns. For almost four decades until 1981, I noted, Congress allowed the Fed to borrow directly from the Fed, so that it might safely maintain a low TGA balance.

The second bit of new information that's come to my attention concerns this second proposal.

Providing for Emergency Treasury Borrowing

It turns out that, back in September 2006, the GAO published a report titled "Backup Funding Options Would Enhance Treasury's Resilience to a Financial Market Disruption." The report was written in response to the 9/11 attacks, with the aim of exploring ways in which the Treasury might continue to fund its expenditures "during a future wide-scale financial market disruption" without having to maintain a large TGA balance. In other words, the report was aimed at helping the Treasury to deal with precisely the sort of problem it chose to address in 2015 by maintaining a mammoth TGA balance. The difference is that back in 2006 the Treasury considered a large TGA balance wasteful. So it encouraged the GAO to look into more efficient options.

One option to which the GAO devoted considerable attention was that of once again allowing the Treasury to borrow directly from the Fed under certain conditions. The report refers frequently to the Fed's former direct draw authority, devoting a lengthy appendix to it. It also notes that "primary dealers and commercial bankers generally agreed" that allowing the Treasury to borrow directly from the Fed "was the most resilient and direct way for the Treasury to ensure it met its obligations," while stating "that they did not think this arrangement would damage the Treasury's or the Federal Reserve's reputation if it was used in a limited way during wide-scale disruptions."

Fed and Treasury officials, on the other hand, worried that the Treasury might abuse its ability to borrow directly from the Fed, particularly by taking advantage of it to evade the debt ceiling. Still, rather than oppose the idea altogether, they settled for suggesting that any direct Treasury borrowing rights would have to be "very tightly" controlled, in part by having to be approved of by both the Fed Chair and the Secretary of the Treasury, if not the President.

To allay such concerns, the GAO concluded that, "Although the existence of a previous draw authority is relevant," it would not be prudent to suggest simply restoring that authority. Instead, it recommended a "two-tiered" approach to emergency Treasury funding. The first tier would consist of lines of credit established by the Treasury with various commercial banks, and of private placements of Treasury CM ("Cash Management") bills. Only if these options proved unavailable would the Treasury have direct recourse to the Fed, making it a truly "last resort" option. To guard against the possible use of that last resort option as a means for evading the debt ceiling, the debt limit could simply be redefined to include any direct Fed lending to the Treasury.

The Direct Treasury Lending Bugbear

As someone wary of excessive central bank power, who by no means assumes that we need never worry about inflation breaking out again, and who favors (for all its limitations) central bank independence, I might be expected to blanch at the prospect of having the Fed allow the Treasury to borrow directly from it. In fact, I'm not inclined to do so, for the simple reason that, widespread opinion to the contrary notwithstanding, so far as the risk of inflation is concerned I don't believe the Treasury's ability to borrow directly from the Fed matters much at all.

The extent of government spending, deficits, and inflation, is indeed a function of the ease with which the Treasury can borrow; and that ease is, in turn, a function of the Fed's interest rate settings. But that said, whether the Fed lends directly to the Treasury, or only purchases Treasury securities on the open market, makes little difference. So long as the rate at which the Fed lends directly to the Treasury is consistent with its general rate targets, the Treasury can't generally do better by borrowing directly from the Fed, or having the Fed buy bonds directly from it, than it might by selling more bonds on the open market. (As for the 1935 decision to prohibit direct Fed purchases of Treasury securities, Kenneth Garbade explains that there has never been any good reason for it.)

What matters isn't whether the Fed funds the Treasury directly or indirectly, but whether the rate at which the Treasury secures funding is or isn't consistent with the Fed's monetary policy objectives. Although the Treasury can benefit by being able to borrow directly from the Fed, its gains consist, not of enhanced opportunities for inflationary finance, but of greater short-run funding flexibility, including the ability to borrow on short notice even when private markets seize-up.

So far as I'm aware, the Treasury has only abused its direct borrowing powers twice, in 1977 and again in 1979, taking advantage of them on both occasions to borrow in anticipation of reaching the debt ceiling. This abuse of direct borrowing, which a more carefully-worded statute (or more carefully-defined debt ceiling) could easily have precluded, was among the factors that informed Congress's decision to end the Treasury's direct draw authority in June 1981.

A Word from Canada

The third and last bit of new information comes from Jean-François Nadeau, Senior Economist for Canada's Department of Finance, by way of David Beckworth. In response to David's December 16th Macro Musings podcast, in which I discussed with him my ideas for limiting the Treasury's use of the TGA (among other topics), Mr. Nadeau sent him an email pointing out that some of my proposals were "already taking place in Canada." The Bank of Canada, he wrote,

achieves the near-zero settlement balance in the LVTS [Large Value Transfer System] by essentially putting Government of Canada balances on auction twice a day so as to offset any payments from LVTS members (essentially commercial banks/credit unions) to the government that would risk creating a shortfall in settlement balances at the end of the day.

Mr. Nadeau also noted

that some of the institutional constraints that seem to plague the US system have been dealt with in Canada. In particular, the Bank of Canada regularly buys Government of Canada bonds on a non-competitive basis on the primary market, what many would call money printing. It even announces ahead of time how much it is going to buy directly.

Finally, he observed that the Bank of Canada

authorized by law to lend money directly to the Government of Canada and even provincial governments on either a secure or non-secured basis, although in the latter case there is a limit.  In essence, this is the same thing as allowing the US treasury or central banks in the foreign repo pool to be counterparties to the standing repo facility that is proposed right now.

Speaking of HM Customs, Laurence Stern famously opined that "they order this matter better in France." For my part, I maintain, regarding the U.S. government's handling of monetary policy both today and more than a century ago, that they order the matter better in Canada. So I'm naturally encouraged by Mr. Nadeau's remarks. The Fed and the U.S. Treasury would be wise to take a good look at Canada's system, and to allow themselves to learn from it.


*I'm grateful to my friend, Alex Schibiuola, an economist at the Treasury, for discovering the amendment and making me aware of it.