Interest Rate Manipulation Comes Back to Haunt Its Most Ardent Supporters
It is difficult to give too much deference to commentators, particularly economists, that speak to the value of quantitative easing in such bland and generic terms. It almost sounds exceedingly easy, as if the Federal Reserve buys a bunch of mortgage bonds, mortgage rates decline, more people can afford to buy houses and the world is full of sunshine again. It usually doesn't get more detailed than that, partly because it is now ironclad law that in order to reach a mass market demands such simplicity, but also partly because that is the extent and depth of the profession's actual knowledge of the inner workings.
I hear it all the time when these same persons, who are nearly monolithic in their undying devotion to the sanctity of the FOMC, try to describe something so basic as the "money supply." Such a notion in the 21st century is so entirely fungible as to lose all proper and tangible meaning, yet that doesn't stop a considerable number of respected commentators from removing all real world complexity.
For example, Deutsche Bank announced last weekend a new "capital" campaign whereby the bank will raise €8 billion in order to focus more on the high yield and leveraged debt markets in the United States (this is not a joke). Part of the reason is that those debt markets are where all the action is now (thanks to what, exactly?) given that FICC trading, boring fixed income and bond trading, is almost completely dead today - the very segment that had sustained the global banking enterprise from the depths of near total despair.
How will Deutsche Bank get its new euro capital from Frankfurt to NYC? It's all the more amazing since 60 million shares are to be (have been) sold to the investment company of a (the?) Qatari Sheikh. There will be a wonderful trip through derivatives markets and bank balance sheets (no doubt including Deutsche's London and US subs), requiring the accounting and finance acumen of dozens of systems including risk management. Do we include or exclude the Sheikh's initial funding toward the US$ money supply?
More interesting is why FICC trading has become so unprofitable. The short answer is the very same people who thought interest rate manipulation was a terrific idea in the first place. But such a generic statement plays right into the very critique I offered at the outset. There is obviously, given this setup, much more "nuance" and "texture" to how policy built up FICC only to tear it apart and set the world of finance into a much more unsettled position.
First, you have to realize that the Fed through FRBNY's Open Market desk doesn't just buy some mortgage bonds as if they were like US treasuries. QE actually operates deep within the bowels of mortgage bond trading in a place called TBA. The entire purpose of the TBA market is to provide liquidity to something that is, at its core, completely and totally illiquid. A mortgage loan is about as static as it gets in banking.
What the Fed is buying through the Open Market Desk are largely "production coupons." The TBA market is a highly standardized operation allowing millions of individual mortgage loans to be packaged into MBS securities in such a fashion that these otherwise immovable loans can be turned to cash in a moment's notice. But mortgage originators need to "buy" GSE guarantees and factor that cost into the setting of MBS prices (along with a set aside for servicing costs). So whatever the net yield on the mortgage pool, say for argument's sake it is 5%, the originator will pay 50 bps to the GSE for its guarantee, set aside 25 bps for servicing costs and then subtract its own spread. If that profit spread is 25 bps, the "production coupon" that is left is 4% to the market.
The Open Market Desk's purchase of production coupons amounts to a retail purchase out of what is really a wholesale product. The generic ideal is that by purchasing more production coupons than might have otherwise been bought it will allow more room for originators to pocket a spread. In other words, if the Fed purchases bump up demand to the point that the "market" is competing for production coupons at 3.75% instead of 4%, the originator can gain some additional bps in profit spread and even pass some of those savings to new mortgage loans in the form of lower interest rates. The increased spread should, theoretically, entice more participation and increase production of mortgage loans to add to the TBA pool (because there is more profit to be had).
That amounts to an artificial subsidy to this type of finance, meaning any increase in activity is done so because of this sponsorship rather than the fundamentals of actual mortgage and real estate conditions. The Fed wants what it wants, including and especially an artificial "pump priming" process that it believes (wrongly, as it is turning out) will restart the housing market.
There doesn't seem anything too evidently amiss by this QE process as I'm sure there is no surprise that the Fed is "greasing the wheels" of finance. But there are costs to such grease, some that are only being discussed now in the deep recesses of global banking.
The danger of Open Market operations to such a scale in the TBA market coincides with its position as a futures/forward operation. The actual purchase of production coupons is not immediate, but upon settlement that may be several months into the future. This framework helps originators because they can "lock in" financial factors without having to take on risk of conditions changing between the filing of a mortgage application and the funding of a mortgage loan. That upside to originators poses a bit of a problem in that the pipeline is susceptible to large swings.
The slight rise in mortgage rates last year unleashed a collapse in mortgage lending - the pipeline dried up. To say it was unanticipated by the "market", especially policymakers, is an understatement (yet again). So when the Open Market Desk went about setting its policy-related purchases, it was buying ahead of that collapse, setting up what looked like a liquidity event. Because the pipeline was ultimately much lower upon settlement than at first believed, there was far less supply available to the market for liquidity.
There is actually even more nuance to consider. For some MBS issues, particularly those of benchmarks, the pipeline/gross issuance problem can be reduced because there is a large reservoir of loans that can be delivered into existing TBA positions should issuance conditions become unstable. That would mean pricing in those benchmarks would see little disruption or volatility. For those production coupons without deep pools, an issuance disruption would likely lead to premiums and stratification of pricing - fragmentation.
But the Open Market Desk has the ability to adjust its pace and program based on implied conditions in TBA. If it senses a tightening of supply/issuance relative to settlement, they can, according to Treasury Market Practices Group Best Practices, sell dollar rolls. While that sounds like a circularized stack of Federal Reserve Notes with a rubber band wrapped tightly around, a dollar roll is just another (why not) derivative contract. Selling dollar rolls effectively postpones settlement of any expected deliveries into the production coupon security.
These dollar rolls function much like the securities lending program out of the SOMA portfolio. In rough terms, as the Open Market Desk takes these securities out of the "market" through its "purchase" operations, it has the ability to open a valve and allow them to flow back in. This is, however, by no means a perfect substitute, as you might imagine. That is an important concept.
Dollar rolls themselves are used as hedging instruments by insurance companies, pension funds or any bank with significant rate risk. A new supply of dollar rolls from the Fed as an offset to pipeline problems in TBA is not a perfect substitute of "market" conditions for hedging and handling changes in structure and conditions. In neither setup is the state of collateral in the market the same before QE and after. It sounds, on the surface, as if the Fed has engineered sufficient workarounds to address any significant shortfalls in operations created by its huge displacement, but the potential remains for disruptions leading to a tiered or fragmented structure. Into those conditions enters uncertainty and hesitation.
We don't have to use future tense on that assessment, as in this might occur down the road. I think we have already seen exactly that close up last summer, though no one seems to care much outside of the nervous gatherings and near-frenzied speculations from inside the system. It's not just MBS mechanics that are under scrutiny, as attention has shifted to the corporate space.
As stated before, it was only the threat to disrupt QE that unleashed what can only be termed an incongruous "market" disruption. The MBS market, in particular, suffered a serious rout in June 2013. Part of that was due to leverage shifting from banks to nonbanks, especially mortgage REIT's in the wake of regulatory winds shifting post-crisis, but a full part also seems to be due to what is increasingly described as an unstable liquidity system.
It has become more and more apparent that dealers in these credit markets have removed a significant amount of balance sheet capacity to absorb risk. To create liquid credit market conditions requires banks to provide warehouse services with which they can add and remove, repo and rehypothecate issues in MBS, UST, corporates and even equities. The dealers act as a shock absorber on the system. If, for example, an MBS in TBA began to trade significantly lower because of whatever reason, at some point a dealer should step in much the same fashion as a market maker in stocks.
The dealer's capacity to engage as that liquidity arbiter is very different now than it was before the crisis. For the most part, discussions about unstable conditions are centered around the fact that dealer inventories in many categories are far below where they were before 2008. In corporate bonds, the Fed itself estimated that dealer inventories have been reduced by 70% (though other observers peg that around 40%; either way, it's a significant reduction in capacity).
True market liquidity is difficult to actually define because it is a shifty proposition. We mostly think of it as both volume and the bid-ask spread. Volume gives us an idea of depth, while tight spreads show that both sides of the transactions are sufficiently satisfied as to not demand some atypical premium. The primary problem with these features as liquidity concepts is that they don't tell you about what liquidity will be like tomorrow; only what it was like yesterday and recently. Market participants infer that tomorrow will look like yesterday, but that recency bias is often dangerous (a lesson we keep failing to hold dear enough despite numerous episodes).
By early May 2013, daily trading volume in MBS was down from where it was averaging for the last half of 2012, but not too much as to indicate problems (around $60 billion, which was about even with early 2012 and a great deal more than 2011). The bid-ask spread on MBS in the TBA market was about as "normal" as you could possibly get. The range on the spread had been contained within 3-6bps for an entire year. By most accounts, despite non-specific concerns about QE removing too much collateral (and thereby responsible for some of the decrease in trading volume) MBS liquidity seemed relatively benign and stable.
After the idea of taper was unleashed by Ben Bernanke and several FOMC members that May, trading volume actually rose precipitously, right back up to $70 billion per day. The bid-ask spread, though, spiked all the way up to 10 bps by mid-June, right around the time the MBS market was in a state of complete disarray (which ultimately led to the disaster in mortgage lending that is now seemingly erasing most of the housing bubble of the past two years or so). Much like fractals or the snowball of error terms in statistical modeling, such a small start out of relatively and seemingly favorable "market" conditions led to someplace totally opposite and unexpected. The fractures in the market were just lightly teased out prior to the event to those that were paying close attention, but were well-hidden especially to those whose understanding of QE is limited to bland generalities (which I keep concluding includes most, if not all, policymakers).
There is another facet to consider here too. Dealer inventories have been declining at exactly the same time issuance in these other markets is rising, meaning in percentage terms banks share of corporate credit, for example, is declining. That means that the liquidity absorber function is not just apparently impaired in absolute terms, but very much so as a proportion of where the market is going.
As corporates garner the most concern, there is still something very representative to me about what is going on there. Mortgage REIT's and retail mutual funds and ETF's can't get enough credit issuance, but there are internal plumbing considerations that are going to be at work if (more likely when) the inflection is reached - the difference there is between orderly withdrawal and disorderly rout. When the margin calls begin and feed on themselves, who will be there to catch the falling knife. Or, more precisely, how much capacity will exist to engage enormous risk.
According to the US Treasury last year, "net" issuance of all financial securities has fallen from about $4 trillion on the eve of panic to only $1 trillion in 2013. Gross issuance is partly to blame, mostly as a consequence of the big housing bubble collapse, dropping from about $5 trillion to $3 trillion. The remaining $1 trillion is central bank programs like QE.
That leaves a marketplace with a much-reduced supply of collateral, but as I have tried to show above, it's not just the amount that matters. I would argue that it is the methods that matter as much if not more. In my analysis, dealer willingness toward warehousing risk is being negatively affected by all of this (and tightening leverage restraints, Basel III, Volcker Rule noises, etc.). The lack of net issuance of collateral makes inventories more expensive and less certain. That feeds into risk management as well, particularly given the post-selloff changes in expected volatility.
The effectiveness of hedging is also a primary consideration. The Open Market Desk may be convinced that dollar rolls offer no significant intrusion, but given the behavior in interest rate swaps and even the eurodollar curve I am far less inclined to go along with that interpretation.
To bring this back toward the generalities, it appears to me, and many others, that the credit markets are much less stable now than if QE had never been proposed. Furthermore, I think we got a small taste of that last summer when what should have been a relatively innocuous and minor potential for change in monetary policy unleashed a wave of disruption that rippled not just violently in credit markets but globally (just ask India and Brazil, Indonesia and even Russia). That's not exactly the same interpretation as those offering the artless and crude assessment that the Fed "lowered mortgage rates" leading to a housing renaissance, thereby helping the economy recover. There is also the lack of recovery to go along with this potential instability on the other side of the ledger.