Short answer I don’t know. But it’s worth considering.
As it stands Barclays’ ETN losses are being talked about all over the shop. Bloomberg‘s Matt Levine suspects they’re down to a clerical error by a mid-level securities lawyer who forgot to tick a box in a regulatory filing. Others like new Alphaville editor Robin Wigglesworth call it a simple Snafu.
I, however, wonder if the true cause of the Barclays ETN blow-out could be down to a fairly common practice within the ETP industry, known as create-to-lend, which routinely transfers a large amount of risk onto the books of ETP issuers in somewhat obscure and poorly understood ways.
For those who do not know, ETP is an umbrella term for all products issued via the exchange-traded mechanism. Some are funds. Some are commodities. And some, like the Barclays securities mentioned above, are notes. The latter are possibly the riskiest because of how issuers manage the risk associated with issuing the notes: opaquely.
This story pertains to the suspension of two specific Barclays ETNs, the iPath Pure Beta Crude Oil ETN and the iPath Series B S&P 500 VIX short-term futures ETN.
But before we begin, for the historical record, here is the Barclays share price as it stood Monday:
It should be stressed that the following commentary is entirely speculative and based mainly on legacy knowledge acquired as a result of many years worth of reporting on the topic. When I have more time, I will definitely make some specific follow-up inquiries. But for now, I thought it worthwhile to lay out what I know, not least because, when it comes to the mysterious create-to-lend dynamic, nobody ever wants to talk about it anyway.
But first, here’s what I noted about the Barclays saga on March 16 (my emphasis):
It’s always around this point in the Exchange Traded Product fallout narrative that the mainstream press gets openly confused by the structural complexities of the products they usually like to claim are super simple and retail friendly.
The truth is ETPs are not super simple. They’re highly complex and varied financial structrues. And there isn’t really a go-to standard in terms of what qualifies as an ETP. It really pays to read the prospectuses. Some are collateral backed, others use total-return swaps to synthesise specific exposures. In the case of ETNs, such as here, the positions are entirely internalised by the issuers. In this case that’s Barclays.
What’s happening with the suspension of creations in these products is a direct echo of what happened to the de facto Credit Suisse sponsored TVIX ETN in 2012.
In short, the fact the bank is no longer prepared to internalise long positions in these products means it can no longer shift the opposite position (in this case the short oil and short VIX positions) off its books, and is running up against risk limits in terms of how much it can directly internalise.
How much of this reflects stress at Barclays is a function of what those risk limits are, which is highly proprietary information. For now, the fallout will impact those who shorted the ETN, since the lack of creations is likely to widen the NAV/Price disconnect and make it impossible to buy back shares at fair value.
At the time, few commentators appreciated the scale of the bank-wide exposure that Barclays was carrying as a result of issuing these securities. Nor did they appreciate that this was likely an internal hedging issue gone awry.
A regulatory filing from Barclays on Monday, however, has revealed that the bank will be materialising a £450m loss net of tax due to a mismatch between registered and issued securities.
According to the filing, the bank under-registered securities by a massive $15.2bn (relative to $20.8bn properly registered securities). The filing also notes that the bank has commissioned an independent review to get to the bottom of the incident.
The filing is light on any other details, but if you look closely at the wording there are clues to what might be going on (my emphasis):
As part of its structured products business, Barclays Bank PLC (“BBPLC”), a subsidiary of Barclays PLC (“BPLC”), is a frequent issuer of structured notes and exchange traded notes in the United States and elsewhere. These securities are often issued to meet actual and anticipated client demand for such securities.
In the first instance one has to analyse what does it mean to issue securities in response to “anticipated” demand as well as “actual”? At least in the context of ETNs.
It is fairly normal to pre-register securities as part of programmes that anticipate future demand. This saves on paperwork when demand actually manifests. Is it normal to pre-issue securities in anticipation of demand? Again, in some specific cases yes too. Most often in the context of IPOs or greenshoes.
A greenshoe is probably the most useful metaphor to use to understand what’s going on, not least because ETPs by definition provide the market with a type of perpetual greenshoe stabilisation mechanism in what amounts to an ongoing “offering” of listed securities. A perpetually stabilised IPO if you wish.
What few people appreciate — even in the context of greenshoes — is that this sort of over-issuance in anticipation of future demand amounts to a type of shorting (in which, arguably, the future demand referred to is your own short covering).
It’s a point that used to drive my old boss routinely nuts whenever greenshoes hit the news agenda. This is because time after time the market coverage would inevitably fail to understand the nature of the shorting at hand, balls up the explanation or treat it as some sort of industry conspiracy.
As Paul Murphy once noted with respect to the Facebook IPO in 2012 (yes it’s hard to believe Facebook has only been public 10 years!):
Quick explainer: This price support stabilisation system works by the banks allotting too many shares initially to clients buying into the IPO. If they have to support the market price when trading gets underway by buying stock, the banks are essentially covering their short position. On the other hand, if the IPO goes smoothly (price is up) they exercise an option for extra stock from the issue to cover the initial over-allotment.
An even bigger and more recent debacle involving bad reporting of greenshoes pertained to the Deliveroo IPO. A good explanation about how that particular over-allotment option worked, however, can be found here.
In that example, the stabilisation agent borrowed stock from a third party VC to over-sell into the IPO meaning if the price fell below the target rate the stabilisation agent could move to buyback the stock to cover his short – thus supporting the price. If on the other hand the share price popped, the agent could have used their exclusive over allotment right (the ability to buy share at the issue price irrespective of market prices) to cover the short without any incurring any losses.
What Barclays is describing in its ETN filing is a very similar process. It’s just that in ETP land it’s known as a “create to lend” programme and the role of the various parties is slightly different.
According to my understanding “create to lend” works like this: (Though I should caveat that I’ve had to patch a lot of this together myself as the pros tend not to what to discuss the mechanics as they consider them a proprietary secret sauce. So if a pro would like to correct me on the details I would be utterly delighted! We would finally have an official explainer from the industry.)
- The ETP issuer or principle market maker wants the ability to stabilise the NAV of the ETP to the price. In a conventional ETF they would either buy/accept underpriced units from the market and sell over priced underlying in exchange or buy/accept under priced underlying from the market and sell over priced units in exchange.
- With ETNs, however, there is no underlying to deliver – the fair value of the underlying is usually derived from a formulation and anyone market making the product is usually obliged to deliver cash rather than securities, as priced by the index the ETN is tracking. This is because the underlying is effectively a total return swap based linked to the index the ETN is tracking.
- More often than not the primary market maker is also the issuer, and the true “underlying” security that backs the product is a total mystery. It amounts to whatever hedge the ETN issuer feels can best manage the risk being incurred – that being the ability to redeem or create ETN units at a price that tracks the index they are supposed to be linked to.
- Many hedge funds like to use ETNs or ETPs as shorting tools. Market makers and issuers like this because it allows for two-sided markets that help them better achieve price discovery and price stabilisation vis-a-vis NAV.
- ETN issuers often satisfy this demand by creating stock specifically for hedge funds to sell short into the market.
- Due to the permanent arbitrage mechanism that governs ETPs, however, if HF shorting sales depress the price of the units relative to the NAV, these shares are immediately bought back by the issuer.
- That means that as far as the public picture is concerned the trade is a wash. There is no growth in overall outstanding shares. At least not officially.
- The shares borrowed and sold short by the hedge fund exist in a sort of phantom realm – as liabilities of the hedge fund to the issuer, which in theory can be covered at any time because of the issuer’s obligation to deliver enough securities into the market to keep them tracking the index price.
The problem is that in a popular ETN this sort of phantom “over issuance” can get pretty sizable.
Some, such as Andrew Bogan, have worried in the past that the practice amounts to a type of naked shorting that could one day pose a systemic risk to markets.
ETP issuers, however, have historically dismissed such concerns on the basis that it’s not really naked shorting at all because the issuer can create new shares at any point to cover the liabilities it has incurred. In theory there is no constraint on the ability to create new shares.
Except as the Barclays case (and others such as the TVIX affair before it) proves, this might not always be so true. This is especially the case with ETNs.
The key takeaway
If we assume that the vast majority of the ETN securities Barclays issued but never registered amount to these sorts of shorted “phantom shares”, created under the auspices of a create-to-lend programme (and distributed largely to third party hedge funds), the question we should be asking is this: what is actually stopping Barclays from simply “creating” its way out of the problem?
The answer is no doubt some sort of limitation in Barclays’ own internal risk capacity or the rising cost of running hedges that can support liabilities on the long side of the trade. For example, if you’re offering exposure to the oil price via a TRS, you need to deliver that exposure through your own positions or hedges. That means both buying and funding those hedges – which in current market dynamics are getting pricey, especially for banks which service oil traders. This is due to a sudden dearth in hedging activity, which potentially also translates to a dearth of naturally offsetting flow for hedging structured products products with.
When it comes to hedging Vix products, meanwhile, large banks with sizable structured products divisions tend to generate organic volatility exposures via their day to day operations. Often times ETNs are issued to recycle some of that natural exposure into third party products that can act as de facto neutralizing offsets. If market dynamics impact the size and scope of those cheap natural hedges, an issuer might be reluctant to load up on expensive conventional hedges to offset them. It simply might not be economic to do so. Another possibility is that the scope of the natural hedge the bank thought it had was not as far ranging as originally calculated. In that case it may have underperformed to such a degree that no new exposures can be justified.
Which brings us to the following wording in the Barclays release:
This reflects a c.14 bps reduction from the estimated loss and a further c.15 bps reduction due to an increase in risk weighted assets in respect of short- term hedging arrangements designed to manage the risks to Barclays arising out of the rescission offer. The equivalent impact on BBPLC’s solo-consolidated CET1 ratio as at 31 March 2022 is expected to be a reduction of c.23 bps in respect of the estimated loss and c.23 bps in respect the hedging impact. The hedging impacts will reverse on conclusion of the rescission offer.
This to my mind would imply some sort of gross internal hedging miscalculation – probably resulting from an aggregated (or internalised) group exposure that was considered more of a hedge to these ETNs than it really was. The language of the filing implies that this oversight now seems to be being plugged with expensive short-term hedges which are keeping the legacy exposure from growing larger but which will not entirely be closed out until the losses are materialised via the completion of the rescission offer.
The rescission offer in this case is equivalent to an over-allotment style safety net used by stabilisation agents to cover the risk of the stock (which they have pre shorted stock ahead of an IPO) popping. In the conventional over allotment, the agent retains the right to buy the stock to cover the shorts at the issue price. His downside is thus capped as the right amounts to an option.
In a comparable move, Barclays is now allowing eligible purchasers (most likely hedge funds) who potentially shorted the stock (and are now being squeezed to high heaven because of the suspension of creations) to cover those shorts at the price they engaged into the positions at.
What’s clear is that this seems a preferable option to Barclays than creating the shares the market needs to settle these shorts organically.
One reason for that might be that it would take a lot more creations (and hence a lot more hedging cost for Barclays) to bring the securities back in line with NAV/price than simply buying out all the short positions at a break-even rate for the hedge funds and closing down the fund.
Relatedly, a much overlooked aspect to all of this is how it all fits into Barclays’ Delta-One trading activities. This is the department through which a lot of position aggregation – also known as internalisation – is likely to be conducted. While there’s very little public information about how delta-one operates, what we do know is that it tends to be a profit centre at many banks but also a key generator of risk.
According to an article by eFinancialCareers in May 2019:
Bloomberg claims that vestiges of the old structured capital markets business are now to be found in a Barclays business known as Delta-1 Strategic, run by Sadat Mannan, a structured capital markets veteran. Delta-1 Strategic may be generating up to 10% of Barclays’ stock trading revenues and could have made up to $252m in Europe in 2018, says Bloomberg.
Barclays isn’t commenting on the claims, but insiders tell us it’s not just Delta-1 strategic that has taken over some of the business of Barclays’ structured capital markets business, but the broader ‘Equity Solutions Group.’ This is said to be run by the likes of Manann, plus financing specialist David Lohius, leverage solutions and Delta One managing director Florian Huber, and head of the European equity finance business Mark Newton.
Sadat Mannan has has since moved on to Autonomy Capital Research having been seemingly embroiled in a series of exposés related to how ETP strategies often connect to dividend tax avoidance. But chances are very high that his former department at Barclays may have been involved in hedging activity for these sorts of ETN iPath products.
At this stage all I can tell you is that both Mannan and I are part of the following group on Linkedin:
Which would certainly imply a crossover at the very least.
Of course if anyone knows more I’m all ears. Equally, please do correct me on anything I may have got wrong or assumed incorrectly – especially if you have specific insight.
I plan to make further calls on the matter throughout the week. I should also stress that I have not run any of the above past Barclays.