In a fascinating piece of statistical detective work published in the latest BIS Quarterly Review, three of its foremost staff members argue that the use of FX swaps and forwards is hiding a massive quantity of debt because current accounting conventions place it off balance sheet. They estimate that, in respect of US dollar debt, this financial dark matter is worth some USD10.7 trillion. Their conclusion is that FX swaps and forwards should really be accounted for on the balance sheet like repo. Continue reading
There is a widely quoted paper re-use and re-hypothecation by Egemen Eren — Intermediary Funding Liquidity and Rehypothecation as Determinants of Repo Haircuts and Interest Rates (Institute of Global Finance, 2015) — which provides another example of the danger for researchers of disregarding the difference between the re-use of collateral sold in repo (re-selling) and the re-hypothecation of collateral pledged in margin lending (re-pledging). I have gone on about this confusion in several previous blogs (including a Bank of England paper that excited the financial press and one by Issa and Jarnecic).
The consequence of the confusion on this occasion is that the author misinterprets the post-crisis reduction in the recycling of collateral by dealers as an increase in repo haircuts, whereas much of this change would really seem to represent the withdrawal by hedge funds of rights of re-hypothecation on collateral pledged to prime brokers to cover margin lending and derivatives exposures. The confusion is not academic, inasmuch as the paper appears to add weight to a body of literature that tries to identify the pro-cyclicality of haircuts as the source of the Great Financial Crisis. This academic corpus originated with Gorton and Metrick (2010), who extrapolated from a time series of haircuts on exotic collateral from a single unidentified source in the US to make an ambitious inference about the origin of the entire crisis. It is fuelling regulatory interest in extending the imposition of minimum haircuts across the repo market. The consequences of such an initiative for market liquidity could be serious and should not be based on research that is not firmly grounded.
Eren follows Infante (2014) in modelling the intermediation by a dealer between hedge funds supplying collateral and cash lenders supplying collateral but seeks to extend the analysis to show that the demand for funding by the dealer simultaneously determines both repo haircuts and repo rates. Implicitly, given his data, he is also trying to estimate the rates of re-hypothecation and the cost of collateral pledged in margin lending from prime brokers as well as the haircuts on collateral posted against derivatives exposures.
The proposition is that the dealer in the model obtains funding from the spread between the haircuts paid on repos to cash investors and the haircuts received on reverse repos from hedge funds. This ‘haircut spread’ is seen to provide an alternative source of funding to the dealer’s his own cash holdings or fire sales of illiquid assets (all sales of illiquid assets are assumed to be fire sales). As haircuts expose hedge funds to the risk of dealer default, it is argued that an incentive is required to persuade them to agree to repo (and to give rights of re-hypothecation). The incentive proposed is lower repo interest rates.
The model assumes that the haircut spread is determined by the volume of lending available from cash investors and the dealer’s need for funding, while repo rates are determined by the adequacy of the terms offered by the dealer to hedge funds and their outside options (ie alternative sources of funding, including from other dealers). The conclusions are that haircut spreads are narrow when funding is abundant and wide when funding is scarce.
The adequacy of the terms offered by the dealer to hedge funds is modelled by a comparison between:
- the returns from the investments being funded by the hedge funds plus the value of bonds pledged by the hedge funds but not re-hypothecated by the dealer and instead lodged for safekeeping with a custodian less the costs of repo funding; and
- the cost of the outside options available to hedge funds.
On this basis, if hedge funds have a net liability to the dealer at maturity, for a given haircut, repo rates will depend only on hedge fund returns and outside options (what they can afford to pay given their investment return). On the other hand, if the dealer has a net liability to the hedge funds at maturity, repo rates will also depend on the probability of the dealer’s default.
The author makes bold claims for his model, stating that it can provide an explanation for all of the empirical evidence about haircuts and re-hypothecation before, during and after 2008.
Evidence for the model is adduced in the form of the survey of haircuts in June 2007 and June 2009 conducted in 2010 by the BIS Committee on the Global Financial System (CGFS). Dealers are assumed to the ‘prime’ parties in the survey and hedge funds are assumed to be the ‘unrated’ parties. The implied haircut spreads across all securities increased between June 2007 and in June 2009, which seen as being in line with the predictions of the model.
In addition, it is noted that, in the second quarter of 2008, the fair value of collateral that Lehman Brothers was permitted to ‘repledge’ and the fair value of collateral that it did ‘repledge’ were around a half their values in the first quarter and the second quarter of 2008. Moreover, it is implied from financial reporting of collateral received and ‘repledged’ that, in the third quarter of 2008, that Goldman Sachs needed a haircut spread of 7.2% and Morgan Stanley needed a haircut spread of 7%, while in the fourth quarter, Goldman Sachs apparently needed a haircut spread of 11.2% for Goldman Sachs and Morgan 22%, which is taken to support the model prediction of a sharper reduction in lending to hedge funds by riskier Morgan Stanley than by safer Goldman Sachs.
Is the model and evidence credible?
The problem with the conclusions of this paper is that the haircut spread on repos specified by the author and implied from the investment bank data on re-pledging actually includes the withdrawal of rights of re-hypothecation on margin lending and haircuts on derivatives collateral. In practice, margin lending and derivatives are far more important for hedge funds that repo, which tends to be used for transactions with dealers other than prime brokers. But the withdrawal of re-hypothecation rights by hedge funds is ignored in the analysis.
There are other weaknesses in the paper:
- The idea of a reduction in repo rate compensating for the exposure of hedge funds to the increasing haircut spread imposed by increasingly risky dealers is based on the assumption of risk neutrality. In reality, as noted by the CGFS and others, parties responded to the crisis, not so much by raising prices and haircuts as by rationing credit. Academics tend not to realise that counterparty credit risk is the primary risk faced by dealers given that collateral does not provide a perfect hedge.
- On haircuts, the paper offers only ordinal proof (direction of change) rather than cardinal proof (degree of change).
- The model does not ‘pin down repo haircuts and interest rates jointly’. There is no empirical proof of interest rate behaviour. And the repo rate estimated by Eren is not the market rate but an internal hurdle rate for hedge funds.
- Looking at the “volume of lending by cash investors and dealers’ balance sheets as determinants of haircut spreads” is too narrow a perspective. Collateral price volatility is a key determinant. Collateral price volatility and cash supply/demand are arguably both driven by risk perception and aversion. So a linear chain of causation is unrealistically simple.
However, the hypothesis underlying the paper is not without any foundation. US dealers do appear to impose haircut spreads. So the real problem may be that the paper is too US-centric. This is certainly one conclusion suggested by the work of Issa et al (2016). On the basis of far more compelling data, these authors contradict both Eren and Infante by showing that haircut spreads in the Australian repo market actually narrowed during the crisis. In repo markets such as Europe, there is anecdotal evidence of occasional trade-offs between repo rates and haircuts but only in particular circumstances and under normal conditions. There is no evidence of a widespread or strong relationship between haircut spreads and repo rates. The two are fundamentally different. Haircuts primarily reflect collateral quality and thus price volatility, whereas repo rates primarily reflect counterparty credit risk. Both will be affected by supply/demand for funding and competition but in different degrees, given that haircuts are less transparent and more under dealer control, while the repo rate is largely exogenous.
The paper’s conclusion is that, “The financial system is still vulnerable to liquidity dry-ups and wild increases in haircuts. This paper highlights the need for further theoretical and empirical research that will be of interest to both academics and policy makers to develop tools to mitigate risks in the financial system.”
Unfortunately, the assertion is not proved, which ironically makes the recommendation for more research all the more apposite but only if that future research is based on a better understanding of the legal differences between re-sale of repo collateral and re-hypothecation of pledged collateral.
In a blog in May 2015, I tried to explain that, outside the US repo market, it was incorrect to use the term ‘re-hypothecation’ to describe the use of collateral purchased in a repo by its buyer. Such use is usually termed ‘re-use’ but would be better called ‘re-sale’. The crux of the matter is that re-hypothecation is a right to dispose of collateral that can be given to a pledgee by a pledgor but such a gift is at the discretion of the pledgor. In contrast, the re-use of repo collateral is the proprietary right of the buyer of collateral to dispose of that collateral. This fundamental difference arises because pledged collateral remains the property of the pledgor (giver), whereas purchased collateral becomes the property of the buyer (receiver).
Recognising the distinction between re-use and re-hypothecation is important because of the confusion over what happened to the collateral that was re-hypothecated to Lehman Brothers International Europe (LBIE) by its hedge fund clients in 2008 in order to secure margin lending and derivatives exposures. The inability or unwillingness of regulators to distinguish between re-hypothecation and re-use has had serious unintended consequences for the European repo market.
Barking up the wrong tree
But it is not just regulators who have difficulty recognizing or accepting the difference between re-hypothecation and re-use. In an interesting paper on ‘Collateral Reuse as a Direct Funding Mechanism in Repo Markets’ by George Issa and Elvis Jarnecic of the University of Sydney, the authors kindly refer to my blog last year but then summarily dismiss the substance on the grounds that, ‘Although they are technically different…, the terms “collateral reuse” and “rehypothecation” are used interchangeably to refer to the use of posted collateral in a source transaction to secure a separate transaction.’
The authors are in fact correct to say that the terms ‘are used interchangeably’ but that widespread usage is incorrect, as I tried to explain. Their failure to see the difference leads the authors to propose an invalid reason for the decline in the re-use of collateral in the Australian repo during the Great Financial Crisis (GFC).
The authors argue, ‘… the exploratory analysis showed that collateral reuse experienced a fairly sudden and permanent decline at the start of the GFC. This is consistent with the statement that collateral providers increasingly demanded segregation of their collateral during the GFC due to concerns about credit risk and the possibility that their collateral would not be returned…Following the shock to financial intermediaries from late-2007, a larger fraction of collateral providers then demand that their collateral be segregated due to an increase in the risk that collateral cannot be returned when requested. Hence, even though dealers have a greater need to reuse collateral to ease their funding constraint, the segregation constraint leads to an overall sharp decline in the repo rate spread. A likely reason for the subsequent persistence of a low repo rate spread is higher risk aversion – and therefore a higher likelihood of collateral segregation requests – relative to pre-GFC levels: not only is collateral more scarce due to regulatory requirements…but also repo market participants are more concerned about counterparty risk due to continued liquidity pressures in global financial markets… although dealers have a greater demand to rehypothecate during crisis periods, the reuse rate declines dramatically at the start of the GFC, reflecting increased demand by source party repos for collateral segregation.’
By segregation, the authors mean no commingling with the assets of the prime broker and, most importantly, no re-hypothecation.
The problem with this explanation for the decline in re-use is that sellers in a repo cannot control what buyers do or do not do with the collateral. Because legal and beneficial title to the collateral is sold to the buyer, it becomes the buyer’s unencumbered property. That is the very essence of a repo.
There was indeed a demand by clients for collateral to be segregated (and the widespread withdrawal of the permission to re-hypothecate) but this was by hedge fund clients to their prime brokers in respect of collateral pledged under prime brokerage arrangements, where the collateral had been pledged to secure exposures arising from derivatives transactions and margin lending, and where the hedge funds had previously given their prime brokers permission to re-hypothecate. It was not in respect of repo collateral and legally could not have been.
The rest of the paper
However, the mistaken explanation for the decline in re-use in the Australian repo market should not detract from the quality of the research underlying the paper. The authors use a Furfine-type algorithm to detect overnight repos from securities settlement data and compile a transaction time series. They use this time series to test the hypothesis proposed by Infante (2014) and Eren (2014) that repo dealers acting as intermediaries impose a higher haircut on reverse repo than on matched-book repos, ie they pay less cash for collateral through a reverse repo than they receive for that collateral in a matching repo. In this way, collateral re-use provides net funding to intermediating dealers. Importantly, Infante further suggests that the higher haircut imposed on the reverse repo is an incentive on the seller to run in the event of a deterioration in the credit of the dealer. The data compiled by Issa and Jarnecic conclusively rejects this hypothesis!
There may of course be local reasons why haircut spreads are negative in Australia and future studies of other markets may discover positive spreads. But there is an important lesson to be learned: not all repo markets are the same. FRBNY and FSB please take note!
Finally, a point of detail worth noting is that Issa and Jarnecic estimate re-use in the Australian repo market at 3.56%, a similar order of magnitude to the estimate of re-use in the Swiss market by Furher et al (2014).
I recently jokingly suggested, in response to the draft assignments you submitted, that some of you had such poor grammar and punctuation that your only future in finance might be as an internet fraudster. This profession is notable for slack drafting. You may recall that a spelling mistake recently led to the discovery of a cyberfraud on the Fed.
However, I have just read the ‘Discussion Paper on the Draft RTS and ITS under SFTR’ published by the European Securities and Markets Authority (ESMA) on 11 March. Clearly, grammar and drafting were not high priorities in the department that wrote this beauty. Missing verbs and articles, incorrect prepositions and many other solecisms litter this document, which is not helpful, given that it extends to 187 pages and 145 questions. I should also mention the description of a yes or no answer as a Boolean value: correct but unnecessary showing off.
But, dear students, for you, this offers hope. ESMA can be contacted at 103 rue de Grenelle, 75007 Paris, France.
I have been reading through the Basel Committee of Banking Supervisors’ (BCBS) Consultative Document of 5 November 2015 on Haircut Floors for Non-Centrally Cleared Securities Financing Transactions. This sets out the proposed incorporation into Basel III of the recommendations of Workstream 5 of the Financial Stability Board (FSB) on securities financing transactions. In particular, the paper sets out the framework for minimum haircuts to be imposed on non-centrally cleared SFT between banks and non-banks. If transactions do not meet these minimum haircuts, they will be considered uncollateralised for the purposes of capital adequacy calculations. The methodology of the proposals is problematic.
The first problem is that the measure of haircut used by the BCBS is not a haircut at all. A haircut is a discount of the cash value of an SFT to the value of collateral, ie the ratio of (1) the difference between the value of the collateral and cash to (2) the value of the collateral. The BCBS “haircut” is, in fact, a bastardised version of an initial margin. An initial margin is a premium in the value of collateral over the value of cash, ie the ratio of the value of collateral to the value of cash. For example, the BCBS/FSB “haircut” of 6% is actually an initial margin of 106%. The true equivalent haircut would be 5.66% (= (106-100)/106).
While the BCBS are being consistent with the way haircuts are expressed in Basel III for adjustments to collateral and exposure values for capital adequacy calculations, it is confusing to use the word “haircut” in regulations that are supposed to be applied at market level. This criticism may seem semantic but it is worth recalling the problems that the confusion between haircut and initial margin caused with term repo just before GMRA 2011. It would also be a good idea to ensure the BCBS terminology was consistent with the reporting requirements of the EU SFT Regulation (SFTR).
The second problem with the BCBS haircut framework is the method calculation of haircuts for individual SFT within a “netting set” (ie a portfolio of SFT under the same master agreement), at least as set out in the example on p6 of the BCBS paper. Each SFT with cash in the same cash currency or the same type of security against the same type of collateral (where the typology is that of Basel III supervisory haircuts) is given a representative “artificial traded haircut” according to the formula:
where Ci is the sum of the collateral Sj (which can be cash in the same currency or the same type of security) and Ei is the sum of the exposures created by lending the same currency or same type of security Sk. The artificial traded haircut is represented by Hj,k. This means it is a haircut to be applied to Sj when it is collateralising an exposure to a loan of Sk. The artificial traded haircut has to be compared with the floor as shown in the table below or as implied by the following formula:
The artificial traded haircut only has to be calculated for security type Sj if it is a non-government issue and has been received net within the same netting set. On p6 of the BCBS paper, the following portfolio example is given:
In the example, only security A is liable to minimum haircuts and has been net received. Security A is deemed the collateral in all the trades, which means these should be the following types of transaction:
Trade A = reverse repo of security A
Trade B = securities lending of security B v collateral security A
Trade C = securities lending of sovereign security v collateral security A
Trade D = repo of collateral security A
This view of security A as always being the collateral is confirmed by the calculation of the implied floor f as (1 + collateral floor)/(1 + exposure floor), in which A’s floor is in the numerator.
The problem with the BCBS proposal is the formula for the artificial traded haircut. It is complicated and therefore not good regulation. But the complexity is unnecessary. The idea of an artificial traded haircut is to prevent banks using a combination of several SFTs, which individually may not be subject to minimum haircuts, to synthesize a position which would have been subject to haircuts had it been transacted as a single SFT. For example, a repo of cash against corporate bonds would be subject to minimum haircuts (assuming it was also with a non-bank and was not centrally cleared). But a bank could try to replicate this transaction and avoid minimum haircuts with a repo against government bonds and a securities borrowing transaction between the same government bonds and the corporate bond, neither of which might be subject to minimum haircuts. To prevent such regulatory arbitrage, the BCBS applies minimum haircuts to securities lending and borrowing transactions even where there is no cash. It does so by implying an artificial traded haircut to each collateral type at a portfolio level using the formula above, as set out in the example above. If the artificial traded haircut falls below the floor for that type of collateral, all trades in the portfolio against that type of collateral are deemed uncollateralised.
But is the BCBS formula needed? Why not split each securities lending/borrowing (SLB) transaction into a repo and reverse repo, and imply the haircut from these notional transactions. For example, trade B in the BCBS example, 200 of A borrowed against 210 of B, could be seen as a reverse repo of 200 of A against cash, and a repo of 210 of B against cash. Using the regulatory haircut of 6% for A, the cash value of the implied reverse repo should be 200/1.06 = 188.7. If the cash value of the implied repo is assumed to be the same, then the value of B needed to fully collateralise the cash would, using the regulatory haircut for B of 10%, be 207.5 (188.7*1.10). As 210 of B is actually being given, this transaction falls short of the minimum haircut by 2.5.
In the case of the example portfolio, the difference with the BCBS is that trade D falls below the floor, as too much collateral is being given. Consider how this simpler approach would work in the case of the example portfolio:
|Trade A||reverse repo of 100 v 105 of A||given A’s supervisory haircut is 6%, A should be 100 x 106% = 106||so trade A is short of haircut of 1 of A|
|Trade B||securities borrowing of 200 of A v 210 of B||this trade is equivalent to reverse repo of 200 A and a repo of 210 B: given A’s supervisory haircut is 6% and B’s is 10%, the reverse repo of A should be against 200/106% = 188.7 of cash, so the repo of B should be 188.7 of cash v 188.7 x 110% = 207.5 of B||so because trade B is giving 210 of B, it represents a deficit of haircut of 2.5 of B|
|Trade C||securities lending of 85 of sovereign bond v 90 of A||given A’s supervisory haircut is 6%, A should be 85 x 106% = 90.1||so trade C is short of haircut of 0.1 of A|
|Trade D||repo of 20 of cash v 25 of A||given A’s supervisory haircut is 6%, A should be 20 x 106% = 21.2||so trade D is giving 25 of A, it represents a deficit of haircut of 3.8 of A|
In total, the portfolio has haircut deficits of 4.9 of A and 2.5 of B, which is equivalent to 7.6 of B.
Does this approach lose the protection intended by the BCBS approach against regulatory arbitrage? Consider applying haircuts to the net notional trades in the example portfolio (where sovereigns have been deemed to be equivalent to cash):
|Trade 1||reverse repo of 165 v 170 of A||given A’s supervisory haircut is 6%, A should be 165 x 106% = 174.9||so trade A is short of haircut of 4.9 of A|
|Trade 2||securities borrowing of 200 of A v 210 of B||this trade is equivalent to reverse repo of 200 A and a repo of 210 B: given A’s supervisory haircut is 6% and B’s is 10%, the reverse repo of A should be against 200/106% = 188.7 of cash, so the repo of B should be 188.7 of cash v 188.7 x 110% = 207.5 of B||so trade B has a deficit of haircut of 2.5 of B|
In total, the portfolio has haircut deficits of 4.9 of A and 2.5 of B, which is equivalent to 7.6 of B, the same as when haircuts are applied individually. And now consider applying haircuts to the gross positions of each type of collateral in the example portfolio:
|Position 1||outflow of cash & sovereigns of 165||there should be collateral of 165 x 106% = 174.9 of A|
|Position 2||inflow of A of 370||if 174.9 of A is required to collateralise the cash & sovereigns, there is a surplus of 195.1 of A — this is equivalent to 195.1/106% = 184.1 of cash, which is equivalent to 184.1 x 110% = 202.4 of B|
|Position 3||outflow of B of 210|
In total, the portfolio has a haircut deficit of 7.6 of B, the same as in the previous approaches.
Why is the BCBS using such a complex approach?
Financial journalists have a difficult job. This is particularly true when they are commenting on markets. These are complex, subtle and dynamic systems which journalists have to observe from the outside. Their problems are usually compounded by lack any direct experience of markets and a paucity of professional or academic literature to consult. If journalists turn to market sources, they have to be wary of possible conflicts of interest. And one hopes that journalists recognise their own conflicts of interest: they are under pressure to produce sensational stories.
These challenges mean that there is a risk of poorly briefed and confused journalists conjuring up articles that inflict damage on markets by encouraging poorly briefed and confused politicians to impose regulations that undermine the ability of markets to perform their essential functions. The consequences can be very real: higher financing costs to the economy and loss of livelihood by those working in finance. The risk is magnified where a journalist is leveraging the reputation of an influential financial newspaper like the FT.
An article that raised concerns appeared in the FT in November, penned by the respected financial journalist Patrick Jenkins. It was entitled “Regulators’ boost for securities lending has risky implications: banks use ultra-safe bonds as compliance shortcut”. It was a strongly-worded attack on securities lending and its role in collateral transformation, which is the process of acquiring higher-grade collateral by giving lower-grade collateral in a temporary exchange. The firm doing the upgrading will typically be doing so because, in the normal course of its business, it does not the type of collateral required by counterparties such as CCPs.
Jenkins dramatically accuses banks of going to “extravagant lengths” to “window dress” and “prettify” their end-year balance sheets. “What are they trying to hide?” he asks, given that securities lending is “probably best known as a proxy for hedge fund aggression because the short selling of equities relies on trading borrowed stock”. For Jenkins, it is “yet another example of the backdoor transfer of risk out of banks and into other parts of the financial system — namely the lightly regulated realm of shadow banking — which we may one day come to rue”.
And Jenkins argues that the securities lending industry has form when it comes to dodgy deals, as “only a few years ago, trades, particularly in Europe, were spurred by a tax avoidance wheeze…cross-border dividend payments used to attract punitive tax charges, but if equities were swapped with a domestic shareholder, such as a bank, the liability evaporated”.
What is being hidden, according to Jenkins, is that “banks are bringing German Bunds, US Treasuries and UK gilts on to their balance sheets, and getting shot of riskier equities by posting them as collateral”. And this business is booming because “policymakers have unwittingly created a new supply-demand dynamic”.
Jenkins’ new dynamic is “new rules on bank capital, liquidity and the clearing of derivatives transactions”. He argues that “there is a three-way regulatory arbitrage at play” in which:
- “banks are boosting their liquid assets to comply with the new Basel III requirement known as the liquidity coverage ratio.”
- “The second trick is that the exchange of assets can be a boon to capital, with equities that tend to attract higher capital weightings swapped for “risk-free” bonds.”
- And thirdly, “with so much derivatives trading moving to central counterparty clearing, there is increasing demand for high quality assets to be used as collateral…and for that, government bonds — even borrowed ones — avoid punitive haircuts imposed on some equities”.
Jenkins skilfully manages to line up topical buzzwords for his article’s word cloud: hedge funds, shadow banking, tax avoidance and regulatory arbitrage.
So, what is wrong with Jenkins’ article?
He was absolutely correct to say that banks are upgrading their collateral. Banks need a stock of so-called High Quality Liquid Assets (HQLA) to meet their Liquidity Coverage Ratio (LCR). In other words, banks need sufficient HQLA to sell or repo out to raise cash to fund a projected net outflow during a severe 30-day market crisis. But why are collateral upgrade trades a problem? If your underlying business does not always generate enough HQLA and you do not permanently need more, it is entirely logical to borrow it through a collateral swap. And, if you are going to borrow to cover a 30-day period, it is perfectly sensible to borrow for that sort of period. It would be risky to borrow for one day and have to roll over.
As for swapping from equities to government bonds, Jenkins seems to have been misled by the ISLA survey to which he refers. This does not say that 90% of European government bond lending is against equity collateral. Rather, it says 90% of European government bond lending is against non-cash collateral. While the report suggests that “This supports very much the view that borrowers in securing access to HQLA are almost exclusively optimising balance sheet and risk weighted assets by providing other assets, often equities, as collateral in these transactions”, there is no hard evidence.
But, even if lots of equity is being swapped for HQLA, what’s the problem? Equity has little value for meeting the LCR. Common equity only counts as a Level 2B HQLA, which means it is only eligible if the national regulator permits and, even if it is permitted, is hit with a minimum 50% haircut and is subject to a 15% concentration limit. Moreover, most CCPs do not accept equity at all as collateral. If a collateral swap can extract some collateral value from equity holdings, it is sensible for a well-managed bank to consider doing so.
Perhaps the most serious mistake in the FT article is about the impact of securities borrowing on balance sheets and risk-weighted regulatory capital charges. Contrary to what the article says, borrowed securities do not come onto a borrower’s balance sheet and the borrower does not pay the capital charge on borrowed securities. This is because the risk and return on borrowed securities is retained, if only indirectly, by the lender. For the same reason, the collateral given in exchange for borrowed securities does not leave the balance sheet of the borrower and he continues to pay the capital charge on these securities. So the borrower cannot use a collateral swap to reduce his capital charge by borrowing HQLA against lower-grade collateral. Indeed, the capital charge of the borrower will increase given that he is taking risk by giving collateral to the lender. There is nothing being hidden here.
For the record, borrowing securities will also not help improve the other liquidity ratio introduced to complement the LCR, namely, the Net Stable Funding Ratio (NSFR). The NSFR, which is designed to enforce a stable asset/liability structure, basically follows balance sheet treatment, ie borrowed assets do not appear on the balance sheet of the borrower, while the collateral given does not leave the borrower’s balance sheet.
Nor will borrowing securities yield any advantage under the Leverage Ratio, arguably the tightest regulatory constraint on banks.
It is a pity that the FT article skates over the fact that securities loans are over-collateralised and margined, and that lenders are very counterparty-sensitive, which means there is typically little risk to lenders. And then there is the visceral reaction to short-selling. This can be destructive but generally it plays a very desirable role in cooling over-valued assets and puncturing price bubbles.
Finally, there is the attack on tax arbitrage. Has this actually been all so evil? Of course, some tax arbitrage was and is unethical. But what is so fair about punitive tax charges on cross-border investors? This was why there have been the “single market rulings to eliminate cross-border withholding taxes” mentioned by Jenkins. It could be argued that markets, once again, played a key role in mitigating unfair and inefficient burdens imposed by protectionist and discriminatory governments until the politicians were forced to catch up.
Last year, the Bank of England took a leaf out of the FRBNY’s book and launched a blog site for its analysts. The site is called Bank Underground. Notwithstanding this attempt to sound a big edgy and an unfortunate tendency towards silly titles (eg “Izzy Whissy let’s get Vizzy”), there’s good stuff here. But there is also the occasional suspect offering.
One of these is a blog that attracted considerable interest from financial journalists at the end of last year: “What do we know about non-bank interconnectedness” by Zijun Liu and Jamie Coen (25 August 2015). The focus of press interest was the authors’ statement, “there is evidence that some non-bank financial institutions have entered the core of the repo network”.
This proposition chimed with the widespread but so far unsubstantiated expectation that, as new regulation hobbles the ability of banks to act as intermediaries, non-bank financial institutions (NBFI) will take their place. And intermediation by NBFIs would of course be a manifestation of expanding “shadow banking”, the search for which excites financial journalists in much the same way as the search for dark matter affects astrophysicists. The blog on interconnectedness proclaimed that the Bank of England had found evidence of such activity. Cue phone calls from excited journalists.
The blog discussed the results of an analysis of the top 60 exposures of 20 major UK banks and investment firms to other banks, other financials and non-financials through transactions in repo and derivatives. It graphically illustrated that, while NBFIs accounted for only 9 of the 30 institutions at the core of the derivative network, they accounted for 14 out of 34 at the core of the repo network. The authors added that the NBFIs in the repo core were the largest counterparties in the network (and four of the largest five NBFIs were hedge funds). The conclusion was that there is “a risk that hedge funds effectively become part of the interdealer repo market’ and a warning that ‘if that happens, stresses at one hedge fund may spread to the entire repo market”.
One can appreciate why the journalists were getting excited. The problem is that the authors of the blog do not appear to be entirely clear about the difference between the various types of securities financing transaction (SFT). In particular, they made an assumption that repos are the same as margin lending. This critical piece of information is hidden away in parenthesis on page 5 of the blog, where it says “repo includes margin loans”. No it doesn’t! Pledge-based margin lending is a very different beast to sale-based repo and is much more important to hedge funds. Among many other things, margin lending is intermediated by prime brokers, of whom hedge funds try to use more than one (the larger the hedge fund, the more prime brokers). Perhaps a better, albeit less dramatic conclusion, would have been that hedge fund business is well diversified across dealers. And this would be consistent with the observation that “banks’ direct credit exposures to non-banks are currently small”.
Another fundamental objection to the suggestion that hedge funds have penetrated the core of the repo market is that such a conclusion cannot be logically inferred from the data. The statistical “core of the network” does not necessarily coincide with the interdealer market. This is especially true when one is using data that measures net volume. In the European repo market, some 70% of transactions go through CCPs. In addition, much non-CCP business is netted bilaterally and netting impacts interdealer activity far more intensely than dealer-customer business, so net volumes will tend to overstate dealer-customer business.
Interestingly, despite their bold conclusions, the authors bemoan their inability to identify transactions between non-banks due to lack of “hard evidence” and laud the imminence of the transaction-level repo reporting requirements that are due to be introduced in Europe. Was the wish for data on NBFIs father to the thought that there is evidence of disintermediation by shadow banks?
Of course, the enthusiasm of regulators for wider and deeper reporting is in stark contrast to the misgivings of many in the industry. There are considerable doubts as to the value of the reporting requirements, particularly given the poor drafting of some of the relevant regulations and in view of the debacle following the rushed imposition of reporting requirement on derivatives activity. Questions are also being asked as to how regulators will cope with the deluge of data and how will they sensibly interpret the numbers without a working knowledge of the market? The Bank of England’s blog will be taken as evidence in support of that last objection.
A final point of interest in the blog is the alarm that appears to be occasioned by the discovery that NBFIs borrow from banks and that banks are lending to these NBFIs against collateral. To say that “other NBFIs seem to be leveraging up via the repo market” hints at disapproval. Too much borrowing would clearly be a problem but, as noted already, the authors comment that banks’ direct credit exposure to non-banks is in fact quite small. And if banks are going to lend, surely secured lending is commendably prudent?
 I should perhaps declare a conflict of interest here. It does not appear that academics will have access to the detailed datasets.
The about-to-be promulgated EU Securities Financing Transaction Regulation (SFTR) is another in a long line of new financial legislation emanating from Brussels that demonstrates how not to write law.
Drafting with one’s head in the sand
The whole dysfunctional process would appear to start with officials at the European Commission and policy-makers in the European Parliament, few of whom appear to have much, if any, experience of finance and many of whom give the impression of being prejudiced against markets. They (and the member states acting through the Council of Ministers) produce their own draft versions of a Regulation. These three texts will eventually be merged by negotiation to produce an agreed version.
A key problem is the Commission’s refusal to consult with anyone in the market before forging a near-final text. The origin of this head-in-the-sand approach to law-making seems to have been former EU Commissioner Michel Barnier. He reportedly forbade his staff from consulting externally before producing a firm draft in case they were ‘captured’ by dark forces. As a result, it seems that no one in the Commission dares to check their ‘facts’ with anyone in the market, that is, anyone who knows anything about the market.
The formal stage of public consultation is largely a sham. The Commission seems extremely reluctant to accept amendments, perhaps because this would be a loss of face and amendments might complicate the trialogue with the Council and Parliament. But there is also deep suspicion of the market.
The problem of “micro-drafting”
Another problem is that Regulations are currently being written at a level of detail which means the officials at ESMA and EBA, who have to implement the new laws, have little or no flexibility to fit the legislation to reality. When mistakes are discovered in the so-called Level 1 text, there is little that can be done by the supervisory agencies when drafting the Level 2 technical implementation rules. The Commission seems very reluctant to admit to mistakes. If the error is so utterly egregious that no one can disguise it, the supervisors have to agree to ignore the law (eg in the case of mandatory T+2 settlement preventing the trading of forward repo on regulated trading venues). In the case of the CSDR, ESMA may be hoping that a long delay in implementation will reduce the embarrassment for the Commission sufficiently to allow them to countenance a Level 1 amendment.
What went wrong with the SFTR
In the case of the SFTR, the most egregious product of the Barnier blind-drafting process has been the attempt to define a repo and, at the same time, distinguish between a repurchase transaction (classic repo) and a buy/sell-back. In this case, however, matters have apparently been made worse by the intervention of lawyers from all 28 EU member states. 11 of those countries do not have a repo market; and another 5 or 6 have very small markets. The rest have active repo markets but lawyers who seem to know little or nothing about them. So 16-17 of the cooks had no recipe: the rest had never cooked this dish before.
The table below analyses the definitions that the lawyers produced and inserted in Article 3(5) of the SFTR.
|buy-sell back transaction
|repurchase transaction (5b)|
|how is collateral conveyed||counterparty “buys or sells” collateral||counterparty “transfers” collateral|
|collateral||“securities or commodities or guaranteed rights”||“securities or commodities or guaranteed rights relating to title to securities or commodities where that guarantee is issued by a recognised exchange which holds the rights to the securities or commodities and the agreement does not allow a counterparty to transfer or pledge a particular security or commodity to more than one counterparty at one time”|
|equivalent collateral||“securities, commodities or guaranteed rights of the same description”||“them [as above], or substituted securities or commodities of the same description”|
|repurchase price||“a specified price”||“a specified price”|
|repurchase date||“a future date”||“a future date specified, or to be specified, by the transferor”|
|documentation||“such buy-sell back or sell-buy back transactions not being governed by a repurchase agreement or by a reverse repurchase agreement as defined in point 5b”||“governed by an agreement”|
The drafting in Article 3(5) is fundamentally incorrect. In fact, it is so incorrect that it could be argued the SFTR does not cover the repurchase transactions and documented buy/sell-backs which are actually used in the European repo market.
For example, the Regulation talks about “buy” and “sell” for buy/sell-backs but “transfer” for repurchase transactions? Does “transfer” means transfer of title, in which case, why was that word not used for both types of repo? Or did the lawyers’ conclave think that European repurchase transactions were pledge-based, as in the US? Note the use of the word “pledge” in the definition of collateral for a repurchase transaction. And is it significant that “equivalent” is defined as “of the same description”. This definition is only one facet of the GMRA definition (see paragraph 2(t)). Consequently, the question arises as to whether the SFTR’s definition is wide enough to ensure that no equity of redemption applies to the collateral, which would mean it was pledged, not sold? If the authors were under the misapprehension that repurchase transactions are pledge-based in Europe, presumably, actual title transfer repurchase transactions are excluded?
And given that the definition of the repurchase price is a specified price, might the SFTR exclude floating-rate repo, open repo and repos of index-linked and floating-rate collateral, where the repurchase price is not known at the start of the transaction?
Meanwhile, the definition of a buy/sell-back specifies that this repo type is undocumented. The authors seem unaware of the Buy/Sell Back Annex to the GMRA introduced in 1995. If all buy/sell-backs are assumed to be undocumented, are documented buy/sell-backs excluded from the SFTR?
However, even if the poor drafting of the SFTR does not provide an escape from its provisions, the regulation’s failure to recognise the essential fungibility of repurchase transactions and buy/sell-backs may have created hidden pitfalls for the market. Repurchase transactions are already problematic in jurisdictions with under-developed financial legal frameworks and inexperienced courts, which have a tendency to regard this type of repo as borrowing/lending rather than buying/selling, that is pledge-based rather than based on title transfer. Will the confusion underlying the SFTR’s definition amplify such re-characterisation risk?
Other problems may arise in due course from the unnecessarily complicated and obscure quality of the SFTR’s drafting. In the case of repurchase transactions, why is collateral further defined as “guaranteed rights relating to title to securities or commodities where that guarantee is issued by a recognised exchange which holds the rights to the securities or commodities and the agreement does not allow a counterparty to transfer or pledge a particular security or commodity to more than one counterparty at one time”. Do any exchanges hold securities or commodities and issue guarantees?
And why is collateral for a buy/sell-back defined differently, just as “guaranteed rights”?
It is also a mystery why the SFTR specifies an agreement for repurchase transactions that “does not allow a counterparty to transfer or pledge a particular security or commodity to more than one counterparty at one time”? That is “double-dipping”, in other words, fraud!
Finally, there is also curious requirement that the repurchase date for a repurchase transaction has to be specified by the seller rather than agreed between the parties. Why, oh why?