A Lost Decade, and Good Prospects for Another One
As if from some bizarro world given what is taking place now, in October 2013, Republicans in Congress actually attempted to force restraint on the Obama administration. The primary issue was Obamacare. No budget agreement could be reached by September 30, as the Senate controlled by Democrats steadfastly refused to negotiate the Republican House’s proposal to defund provisions for the health insurance law. Without an appropriations resolution in lieu, the result was a federal shutdown lasting 16 days in total.
While all that was going on, US T-bill rates rose. The equivalent yields on Treasury bills had been consistently less than 10 bps ever since the panic. Notable exceptions were at some points in 2011 when the 4-week instruments would sport a negative yield, as well as at the end of July that year when another budgetary element, the debt ceiling, was similarly a possible impediment to normal Treasury function.
On October 11, 2013, the Hong Kong Exchanges and Clearing changed the repo requirements for using US Treasury bill collateral. The primary haircut had always been 1%, an uncontroversial low reduction fitting the instrument. Because of the shutdown closing in on a second week, the exchange decided that it would take further precautions that many of its participants had already instituted. The haircut on US bills with maturities of less than 1 year was raised to 3%.
It’s not that money market participants were afraid the US government would default, but that any principle bill maturity payments delayed as a result of the political drama would be treated as a technical default. Repo doesn’t do default, even by a little.
The 4-week equivalent yield (Treasury bills are priced and paid at a discount to pay; the equivalent yield is a calculation to turn the bill’s current price into an extrapolated interest rate for comparison purposes) on October 1 jumped to 10 bps from 3 bps. It would rise to 27 bps one week into the shutdown, and then as high as 35 bps on its final day. Such volatility made the Hong Kong Exchanges decision appear wise, perhaps even modest.
The reason for haircuts in the first place is volatility. A cash provider, collateral taker is not interested in holding a Treasury bill or any other collateral asset for anything other than security. Thus, its only real consideration for the terms of the repo transaction is how much in price that security might move while the cash provider is holding it. The haircut is the standard view of that anticipated volatility, affording the collateral taker enough margin to be more than reasonably assured the price will not deviate so far as to leave some portion of the loan (a repo is really just a collateralized loan) exposed to loss in the event of any kind of counterparty default.
The rise in bill rates in October 2013 was largely a result of OTC repo markets, which is the vast majority, increasingly adjusting their terms for using T-bills as collateral due not to big risks but changing risks (even a shift from no risk to small risk can have enormous effects when the whole system is calculated for only no risk).
The 3-month bill yield never really moved all that much, practically nothing in comparison to the 4-week variety. There was only a slight upward tick to around 5 bps through the first ten days of the shutdown, and at the highest just 14 bps in equivalent yield on October 15. Thus, we can see where repo market perceptions of changing risks were focused; on the 4-week because it was apparently believed any shutdown would not last 3-months, and therefore technical default wasn’t nearly as likely thirteen weeks into the future as four.
While that situation makes for easy intuition and appreciation, its opposite engenders much confusion. Why, for example, would the 4-week Treasury bill consistently yield less than what is supposed to be a money rate limit? This has happened even during once more changing risk perceptions related to another potential debt ceiling constraint.
This past Wednesday for the first time since the Fed’s last “rate hike” the 4-week bill priced with an equivalent yield above the so-called money rate floor (set by the central bank’s reverse repo, or RRP). During that time, the yield had fallen as low as 24 bps below it.
Some have argued that the drop in yields, including the 3-month bill which briefly priced below RRP, too, is also related to the debt ceiling but in a different way. When it was reinstated on March 16 this year, frozen at that level of qualified outstanding US debt, the Treasury had already undertaken several measures to insure continued smooth functioning.
One of the compromises that came out of the November 2015 negotiations that temporarily suspending the debt ceiling was for the Treasury to target its operating balance at $23 billion for whenever the limit would be reinstated. As a result, given that its cash account had accumulated to as much as $400 billion last fall, the auction of Treasury bills was scaled back significantly. A lack of supply would seem to suggest why bill rates might be rich to something like the RRP.
A BIS paper published in September 2014, however, argues, with considerable evidence, that it isn’t supply functions at all that determine this liquidity premium. Instead author Stefan Nagel of the University of Michigan writes in conclusion, “The evidence in this paper suggests that liquidity premia of near-money assets reflect the opportunity cost of holding money.” The focus of his study was the relationship between short-term interest rates and the opportunity cost of holding near-money substitutes such as T-bills. It was not, I have to point out, related to the role of bills as repo collateral.
The perspective of Nagel’s paper was entirely from the cash side. He even contrasts a T-bill instrument with a term repo in explaining why bill rates will often price below maturity-adjusted GC repo rates; the former being liquid for cash purposes. From the point of view of the cash holder, a term repo is illiquid since you are stuck holding collateral (the exception being something like a gold swap) for the duration.
Thus, Nagel argues that when opportunity in general terms is robust, indicated properly by higher and bond money rates, market participants demand a liquidity premium to be able to liquidate any near-money holdings so as to be immediately able to jump on one of many enticing investment choices. The rates on liquid near-money securities like bills should in that situation remain significant less as interest rates rise.
Is that what we are seeing here? The bond market, as eurodollar futures, says no. Even if the Fed was able to “raise” rates on its revised schedule it figures to only get to around 3% at maximum (and its models are always, always skewed to the most optimistic). In terms of general economic and financial opportunity, that’s a world of difference from the 5.25% Alan Greenspan’s Fed managed in 2006, one that still produced his “conundrum.”
What, then, is a bill liquidity premium possibly suggesting at low money rates, even if no longer zero or close to that level?
On the day Greenspan’s final “rate hike” became effective, June 29, 2006, the 4-week bill was yielding 4.59%; the 3-month bill 4.88%. Those were unusually low rates for both, as yields would rise for the 4-week instrument above 5% by August 2006 and largely remain there until April 2007. It would fall below that level again on April 9, 2007, as the word subprime entered the mainstream and would only once more revisit 5% on July 31.
On August 9, 2007, a day that should be as infamous in the economic world as Black Tuesday 1929 is for the Great Depression, the 4-week bill rate fell to 4.60% from 4.94% the day before. The federal funds target rate was still 5.25%. Contrary to Nagel’s argument, that wasn’t due to opportunity-based liquidity premia, it was instead pure risk adjustment. For point of reference, the effective federal funds rate had risen sharply (for federal funds) to 5.41% on August 9, significantly above the target, to then fall precipitously on August 10 like bills, closing at 4.68%.
In conventional terms, neither of those makes sense as either high opportunity (bill rate) or excess liquidity (effective federal funds below target). Like so many other such anomalies (that aren’t any longer anomalies) such as negative swap spreads, the behavior of these rates only suggests that “something” was very wrong. The significance of Nagel’s paper is in this general context to eliminate supply issues (for bills, or just as much any other near-money substitute) and focus on liquidity considerations.
This was a situation that persisted for almost the entire crisis period, causing a whole lot of confusion even among those who hold themselves out as being experts in economics. Paul Krugman, for example, just a week after Bear Stearns’ failure in March 2008 was still struggling mightily with the T-bill rate.
“But right now Treasury interest rates are much, much lower than the Fed funds rate — around half a percent on both 1-month and 3-month bills. Weirdness like negative rates on repos aside (I’m still trying to wrap my mind around that one), basically the Fed can only drive Treasury rates down by about another half-point — which would still seem to leave Fed funds well above 1%...
“Am I being really stupid here? Or is it possible that the fear factor will soon make it impossible for the Fed even to achieve its target on the interest rate it supposedly controls?”
For economists like Krugman, money is something the Fed does, but the market at that time was telling him and everyone else that wasn’t true. Even through the dense fog of rigid ideology he could see very plainly that “something” was missing from the textbook on monetary behavior.
And it was much, much worse than what he was referring to on March 20. The Fed had three days before instituted an emergency 75 bps “rate cut” in the aftermath of Bear Stearns’ demise (technically a merger). That brought the federal funds target rate down to 2.25%. Yet, in those first few days after the shocking development with Bear, on March 19, 2008, the 4-week bill equivalent yield was all of 25 bps, or exactly 200 bps below the policy target; the 3-month was at 60 bps.
It should be noted that in this week after the shock, total UST repo fails were more than $2 trillion (both “fails to receive” plus “fails to deliver”), rising further to $2.3 trillion the week after that. Those levels would only be surpassed in the weeks after Lehman Brothers.
Even worse for the Fed, the effective federal funds rate was all over the place. On March 20, 2008, the day of Krugman’s lament, it was 2.22% very close to the 2.25% policy target. But the day before as well as the two trading days after, the effective rate averaged out to just 2.08%, which would seem to indicate an excess of funds again. Then on March 25, the effective rate jumped to 2.42%, well above target (tight), falling to 2.09% on Friday the 28th , only to trade Monday the 31st back up to 2.51%.
The high demand for federal bills was as their unchanged haircuts and universal acceptance. Haircuts on other forms of collateral were being adjusted often severely, but worse than that after Bear the mere threat of further and more drastic recalculations (remember volatility is at the heart of all this) meant liquidity demands for bills (as well as other UST securities) had completely disassociated them with all other money rates. The roller coaster in federal funds was the rest of money markets trying to make sense of, and deal with, all this.
According to several sources, repo haircuts on subprime related mortgage bonds which before August 2007 were used fluidly as good-standing repo collateral had been moved to as much as 20% (from 1% or 2%) by early 2008. Even haircuts on prime mortgage MBS were altered, to by the immediate aftermath of Bear around 10%. The Treasury bill rates indicated a bank run, really an interbank run, on usable collateral.
We have to refer repeatedly to the 2008 crisis because we are not yet past the 2008 crisis, though in two weeks we will have marked a full decade since it all changed on August 9, 2007. The events of that panic were like high-energy collisions that for physicists allow them to peer into the unseen world of quantum particles. The breathtaking irregularities that appeared around the worst parts like Bear and Lehman have done for us in monetary economics something similar, performing empirical tests about the primordial but universal state of modern money.
What we found was that collateral matters a great deal, and continues to matter to this day. In a great many ways, the system has never recovered largely because collateral (and relatedly balance sheet capacity) remains scarce. The federal government did its part though it didn’t know it, drastically increasing the deficit and therefore the amount of federal securities including bills and on-the-run notes and bonds. It wasn’t nearly enough, however, to offset and replace pre-crisis collateral conditions that once accepted at ridiculously low haircuts all manner of packaged risky securities.
In many ways, that is the central point here. The 2008 crisis was not really an event, per se, but rather a paradigm shift or a terminal end to the prior framework. There was in repo collateral simply no going back; how could there be? The only way to replace trillions in the MBS collateral pool no longer really available would be to manufacture it from somewhere else. Apart from the US Treasury there was no way of doing so under these conditions, and even the gigantic deficits of the early “recovery” period provided nowhere near enough.
Not for lack of trying, of course, as throughout the post-crisis age the system has attempted various means of inventing new collateral methods. It had become fairly common for firms to engage in collateral swaps, mostly trilateral, where junk corporates (including those issued by EM’s in dollars) were exchanged (for a fee) from firms like insurance companies to rehypothecated UST’s technically owned by bond dealers (and who knows who else). The “rising dollar” period, which wasn’t so kind to junk corporates, was yet another reminder that pristine collateral cannot be conjured through swaps, MBS securitization, or other means.
And that brings us back to opportunity again. A healthy and growing economy does not need to depend exclusively upon the federal government for good, solid collateral. A normally functioning financial system will be able to create it through the natural course of growth. It is the lack of opportunity, or the artificial boost of unrestrained monetary expansion, that leads to these sorts of persisting imbalances or shortages.
It’s not as if the repo market was some brand new funding method developed in the early 2000’s coincident to the housing mania. In the United States, it goes all the way back to WWI and, of all places, the Federal Reserve. War-time tax on interest payments paid from commercial paper inhibited the funding of reserve member banks, leading the brand new central bank to extend credit in repurchase agreement form somewhat beyond the scope of its original mandate just a few years old by then. That was 1917.
In now a century of repo, it has only been the last decade where it is an obvious monetary impediment and therefore economic issue. It is because of the lack of opportunity that it remains so, a sort of chicken and egg situation of self-reinforcing feedbacks: collateral problems contribute (along with other factors, like derivative capacity) to further monetary instability, creating a substantial economic drag which reduces opportunity, perceived and real, and therefore the natural way to create more collateral.
The reoccurrence of these concerns in 2017 aren’t truly related to the debt ceiling except as again by way of being revealed by it. Why might there be renewed pressure on T-bill rates this year? With oil prices in another downtrend since mid-February (because demand once more failed to materialize) that has surely given pause to any dealers or repo (or derivative) counterparties taking in corporate junk for collateral swaps (a huge proportion of high yield and midgrade corporates over the past five or six years has been issued by the oil sector directly or by firms in EM countries exposed to oil prices). Maybe even a few minor haircut adjustments were instituted, perhaps forcing participants to overpay for T-bills.
In August last year, Dr. Krugman proved how little he has learned since March 2008 in chastising others for lack of growth. Writing as usual in the New York Times on what he called the Folly of Prudence, he said, “We’ve been living with low-rate, depression economics for 8 years now — and key players are still acting as if they’ve learned nothing.” Yep. He thinks “austerity” is holding everything back, when in truth it’s all the stuff he didn’t recognize a decade ago and still doesn’t now.
In very general terms, the monetary system continues on unable to meet the modest demands placed on it by even the most modest of growth trends (forgetting recovery from the immense contraction). It, combined with other hidden monetary elements, qualifies in every way for Milton Friedman’s interest rate fallacy. The consistency of T-bills, bond rates, “secular stagnation”, etc., is breathtaking in that despite all that the world economy still lost a decade anyway and remains facing the very real prospects for another one.