Busted auction. Hard to imagine today, if for very different reasons, in March 1920 the Treasury Department found itself without enough bids for its overgrown needs. Intending to raise between $300 and $350 million, princely sums for that age, the 4.75% coupon stapled to the 1-year notes being offered came up woefully inadequate. According to contemporary accounts, purchase orders were submitted for only $201,370,500 even though the government had left the subscription books open for a reported two additional weeks.
The wartime demands of fighting men and destructive material behind it, the Treasury Department still needed to keep up its debt payments incurred as a result of the United States’ entry into the Great War. The bill had been steep; the government up through 1918 had to raise about $21 billion. That was twentytimes the national debt before 1917, and seven times more than what had been borrowed to defeat the Confederate insurrectionists just a half century before (still within living memory).
Fortunately, American participation in the European calamity lasted not quite two years and the total national indebtedness “only” got to around $25.2 billion by the eleventh hour on the eleventh day in the eleventh month of 1918. But that left a decade of the 1920’s whereby the domestic economy had begun moving into it with two problems on its back: the massive credit burden along with a topline tax rate which had been ungently punished from 7% to 77%.
As entry into the conflict became inevitable, politically, scholars and politicians heatedly debated how anyone was going to pay for it. There were extreme positions on both sides, though each realized there could never be polar opposites. Wartime financing would have to be a combination of debt and taxes, the only question remaining was the “optimal” proportion.
Many urged more toward taxation for the country to take its beating, financially as well as in terms of lost lives, right from the start. Surrendering to the view espoused by philosopher David Hume, “[E]ither the nation must destroy public credit, or public credit will destroy the nation.” Was it just to suffer future generations presenting them the costs of failures in the present?
The prevailing view was positively medieval; the best way forward was through just enough taxation to keep up immediate expenditures with no interruption while borrowing the rest, the vast majority, from foreign hands as if it might have been possible in the early 20th century to seek recompence from some far-flung Court of Versailles as the first de facto Treasury Department had at the nation’s founding.
Given that WWI was, by its very name, a world war, there was no lucratively untapped potential creditor potentate whereby the American economy could avoid the necessary reduction to fit the thing under its common means. As influential economist Roy Blakey had put it in January 1918:
“When we stop to think we know that it is not twenty-one billions of dollars which our government wants ultimately, but twenty-one billion dollars worth of commodities and services. Our national income does not consist of forty-five or fifty billions of dollars of gold, silver and paper, but of that many dollars worth of wheat, lumber, minerals, clothing, automobiles, etc. There are less than five billions of actual gold, silver and paper dollars in existence in the United States. These dollars are the counters in terms of which the real things are measured and by means of why they are exchanged more easily.”
But in order to raise that awful sum of $21 and eventually $25 billion, money had to be removed from private hands thereby depriving the domestic system of its measurable monetary lifeblood. It ended up having to be this way simply because taxation even at confiscatory rates would never have been enough, so debt, domestic debt, was going to be a practical choosing of the least-worst option. As Blakey wrote, “Munitions of 1930 and men not yet born cannot be hurled against the enemy’s lines.”
This acknowledgement didn’t stop the ingeniousness by which authorities tackled this great money problem; necessity the mother of invention, here was of the greatest necessity and its proud child a further evolution of ledger money result.
The federal government would raise its borrowed funds through massive Liberty Bond drives. There were five in total, the first conducted between May and June 1917, while the last occurring in April and May 1919 (following war’s end, this final bond was relabeled as the Victory Liberty Loan). This all-big-auction schedule necessarily meant raising billions in certain months while expenditures paid out incrementally over the intervening six months or so between bond sales.
To minimize the direct monetary impact of these large clusters of government financial footprints, conspiring with the Federal Reserve, both state and national banks were allowed to simply credit the government for any bonds purchases on its own account as well as on account for its customers. Called War Loan Deposit Accounts, depositories needed only demonstrate sufficient eligible collateral and capital to avoid having to actually pay cash (by transfer to the local Federal Reserve Branch).
In other words, under normal circumstances, a bank customer wishing to invest in Liberty Bonds would have had the funds immediately deducted from their account, credited to the bank, which then forwarded the cash from its reserves to Treasury’s agent, the Federal Reserve branch. The drain of real money and reserves for each government bond sale would be immediate, and since these were herded together in the few huge auctions they would inevitably create highly problematic bottlenecks – monetary shortages.
However, since the government didn’t need to use most of the funds raised by Liberty Bond sales immediately, by crediting the Treasury Department’s War Loan Deposit Account instead of front-loaded cash transfer, the local bank would be able to create a paper liability without the lumpy real money withdrawals. The feds would, over time, convert their deposits from the banking system as a whole to better match the timing of expenditures.
The net result was a truly minimized negative imprint; for the banking system altogether, when Treasury did draw down on its paper deposits, the flow of cash and reserves to the Fed was almost immediately put back into the banking system by the government spending it right back into the economy. Individual banks could better manage their liquidity needs, and while they might not individually get back deposits claimed by Treasury, the banking system as a whole would be beneficiary of its return payments.
Not quite fully developed, yet a cute if primitive “money-financed fiscal expansion.”
It wasn’t without its problematic quirks. One such was the mismatch between what banks would pay Treasury for funds held in deposit on its behalf and what the market might offer for other money instruments. The War Loan Deposit Account was set at a flat, unmoved 2% rate; meaning that private banks crediting the government with a paper liability would pay the Treasury a 2% annual cost for the scheme.
Functionally, this had meant that the US government became a relatively stable, at least predictable funding source for private banks. Knowing that actual deposit or cash transfers wouldn’t become necessary until weeks if not months after any bond sales – and this arrangement survived for more than a decade after Liberty Bonds – they’d deploy the cash and reserves left in their hands (after customers paid them cash) toward more profitable uses; including paying down “borrowed reserve” balances incurred at the Fed’s twelve branch Discount Windows given much higher discount rates.
As a direct consequence, banks began bidding well in excess for whatever the Treasury Department might auction so as to obtain this cheap paper funding. So much so, that the depository system in general had been willing to pay up for government bonds far in excess of how much the government underpriced these same instruments. Bond auctions typically were oversubscribed with bank bids far beyond the par amounts being sold.
March 1920, however, was an exception. Exceedingly large tax payments (draining systemic cash) along with other factors had left the government exposed to a squeeze when trying to issue debt in relatively infrequent, large helpings. Treasury’s annual report for 1920 recalls:
“With this relatively light subscription and the increasingly heavy burdens resulting from payments to the railroads under the transportation act, it soon became evident that the most sanguine expectations of February 12 [1920] would not be realized, and that it was necessary, in order to meet the enlarged requirements of the Treasury and insure the absorption of the securities by investors, to diversify maturities, to introduce loan certificates in moderate amounts, and to offer certificates at rates of interest in line with the market rates for money.”
The side effect of doing so only aided in the carry potential of banks oversubscribing these more frequent auctions still in demand of the 2% cheap government funding. Over the rest of the twenties, more auctions were conducted but with mostly (sometimes only) the banking system participating in them.
Essentially, a two-tiered market; the primary offering price was the par value set by the government, in which the banking system way overbid, while the secondary market price was at times substantiallydiscounted to it. In other words, banks would willingly seek out auctioned government bonds and pay a premium to do so, knowing full well they’d take a loss when eventually selling the instrument to an investor in the secondary market.
According to other contemporary accounts, by 1929 it had long been standard practices pre-selling Treasury certificates of indebtedness (a short-term coupon bond) at a loss of two- and sometimes four- or six-thirty-seconds to still make out awfully well on the entire leveraged trade – provided that government paper liabilities would be maintained at a sufficient level for a sufficient amount of time in order to take advantage of the War Loan Deposit privilege.
In one key sense, Treasury debt had become something other than an investment for banks, rather a pretty specific balance sheet tool by which the depository would secure specific funding parameters that allowed it to best leverage its own situation.
This is not, strictly speaking, a direct comparison to modern repo, yet the outlines of this later wholesale technique cannot be denied when reviewing what seems an ancient parallel. An intermediated method to obtain preferential bank funding with government securities situated in that middle, thereby driving some significant portion of demand, and an unmistakable material impact on price, for those securities for reasons having almost nothing to do with the credit or investment considerations otherwise embedded within them.
We’ve heard time and again since early 2018 (going back to 2016, according to one popular strategist at Credit Suisse), how there are, and must be, “too many” Treasuries. First, the Tax Cut and Jobs Act of December 2017 greatly expanded the deficit necessitating more auctions while more auctioned at the auctions. Then March 2020’s CARES Act simply obliterated all prior views on what counts for the fiscally grotesque.
Predictably, “too many” Treasuries has again resurfaced in its wake, the warning cry of those clung fast to the “bond bubble” viewpoint. While it is true that the US government is the absolute broke-est institution human deviousness has ever conceived, it does not necessarily follow that further action along those same lines must represent “the” step going too far.
If one is to reconcile these two points, supply with demand, a good place to start is by appreciating other potential – and potentially valuable – utilities of the instruments created by federal government recklessness and the long history behind them.
To that end, despite a massive increase in auction size, just this week the Treasury Department sold $60 billion (to the public – largely via, of course, the banking system’s primary dealers) of 2-year notes at record low yields. That’s 50% more than what had been standard during 2019 and the first few months of 2020, and nearly double the par from back before the debt “deluge” early in 2018.
In the bill market, yesterday demand was so hot for these hugely liquid, always-OTR securities that 4-week (28 day) paper was so oversubscribed the high auction price ended up being 99.995722 (since bills are sold without a coupon payment, the government gives buyers a discount in price to par which is then converted into an equivalent interest rate), an equivalent yield of 5.5 bps!
Not only that, only a tiny bit more than a third was allotted at this “high”; the median auction yield came in at 3.5 bps, with a low (5% of accepted bids) practically zero (1.5 bps).
Throughout this 21st century epoch of “too many” Treasuries, the recurrently unquenched emotional impulse which actually dates back to 2008 and 2009, there isn’t even a hint for a busted auction. Primary dealers, those who are statutorily required to bid if no one else does, continue to willingly absorb whatever the government will issue. And then some.
The question really is, as it has been, why?
When you realize UST’s, especially bills, are mostly balance sheet tools, of a pretty specific type and not just for primary dealers (the broader, systemic implications of “securities lending”), the question answers itself. The Federal Reserve does bank reserves and calls it money as if there’s nothing else to it (or as these might be relevant forms of it). Going all the way back to World War I, barely a few years into this central bank’s existence, to put it mildly, it’s always ever been a bit more complicated than that.
