eSecLending rides the asset safety wave
The principal risk in securities lending is that the assets lent are not returned by the borrower. This is why the custodian banks which intermediate the loans indemnify their clients against the loss of their securities, and the borrowers collateralise the loans to make doubly sure. But collateral creates further risks of its own. The sum realised from the sale of the collateral may be insufficient to cover the loss of the securities. And if the collateral takes the form of cash, it creates reinvestment risk, which is not indemnified by the custodian banks as agent lender.
These risks came vividly to life in the financial crisis. Collateral valuations proved optimistic, and collateral markets illiquid. A number of cash collateral reinvestment pools suffered losses. These were borne mostly by lenders, though some lending agents took the view that it was better to make clients whole, even though they were not indemnified. In cases where continuing to lend meant losses were not crystallised, agent lenders restricted withdrawals from programmes. Clients that insisted on retrieving cash collateral were in some cases paid in specie rather than cash.
In an industry not noted for the length of its memory, this experience was sufficiently traumatic to prompt some institutional investors and fund managers to withdraw from securities lending altogether. Others preferred to continue to lend, but to identify and use additional methods of managing and mitigating the risks. The segregation of assets lent, and of collateral received, was an obvious way to do that.
Segregation offers lenders transparency into which of their assets are on loan and who to, and what securities they hold as collateral. It also gives them control over how and where their cash collateral is reinvested. Segregation of assets has enjoyed regulatory encouragement too - especially in Europe, through measures such as the European Market Infrastructure Regulation (EMIR) and the Alternative Investment Fund Managers Directive (AIFMD).
For eSecLending, the independent third party securities lending agent, this renewed interest in segregation has provided an unsought endorsement of their business model. After all, eSecLending was founded in 2000 in the belief that the segregation of a securities lending programme would enhance performance, chiefly by avoiding the cross-subsidisation of lenders inherent to the classic lending pool programmes run by global custodian banks in their capacity of agent lender.
“Performance is always correlated to the underlying assets of the fund, and the parameters it sets,” explains Simon Lee, managing director, business development, EMEA, at eSecLending in London. “In a custodial programme, your performance is correlated to the other lenders in the pool, the parameters you set relative to those other lenders, your position in the queue as opportunities to lend arise, and the algorithm the custodian uses to allocate those opportunities – all of which dilute individual lender performance.”
The lending portfolios of clients of eSecLending, by contrast, are managed on a segregated, stand-alone basis. “In investment management, pension funds have long insisted that asset managers run separate accounts for them,” says Chris Jaynes, president, head of client relationship management and business development, and one of the founding members of the firm. “Yet in the securities lending world, all you could do was enter a commingled fund. You had no option to run a separate account, where you were not impacted by the strategy of others.”
The pooling of client assets reflected the origins of securities lending in the banking industry (in which the transfer of ownership to the borrower and the operation of omnibus accounts are the norm) rather than the investment advisory business. But the decision of the founders of eSecLending to break with the bank-driven model and replace it with an agency function had nothing to do with making securities lending safer. It was about lifting performance.
“The message I use with clients is that `We manage your own programme, you do not participate in ours,’” explains Lee. “Corny as it sounds, that is how it works. Each client has different assets, and different views and goals. It makes no sense to throw them all into a great big pool and expect to come up with the right answer for any of them. Each client has their own securities lending programme, which we manage on their behalf. There is no pooling and therefore no cross-subsidisation of the returns of other members of the pool.”
Nothing chimes better with the spirit of the age in financial markets than the absence of the pooling of risks and rewards. The segregation of assets at every level in the securities chain – global custodian, sub-custodian, executing broker, central securities depository (CSD), clearing broker, central counterparty clearing house (CCP) – has become the preferred option of institutional investors and fund managers as well as regulators.
A firm founded on the principle that large clients with valuable and diversified portfolios are bound to be disadvantaged in a pooled securities lending programme, in which the returns are allocated to members of the pool in an opaque fashion, is finding clients as interested in its ancillary benefits - safety, control and transparency - as performance.
Greater transparency into the size and sources of return gives lenders a higher degree of control over counterparties and the management of the collateral they receive. This is of singular importance, given what happened to classic agent lending programmes in 2008, when cash collateral was pooled with that of other investors and invested in short term money market instruments which proved illiquid or even went into default.
“Each client’s collateral is managed in segregated, stand-alone accounts,” explains Lee. “If we have 10 lenders with exposure to a single investment bank, we have 10 separate mark-to-market conversations that day, with potentially 10 collateral movements. A custodian with 10 lenders exposed to the same investment bank would have one mark-to-market conversation and would allocate the collateral on their own books, at the lender level.”
Unlike an increasing proportion of the revenues of some agent lenders, eSecLending never acts as principal in any securities lending transaction either. Nor does it take assets into custody. The assets of its clients remain with their custodian bank – just in a segregated rather than a pooled account. The custodian also hosts the segregated account which holds collateral received from borrowers. But it is eSecLending which moves securities (to borrowers) and collateral (from borrowers) between these accounts.
“We are instructing on all loan transactions to and from custodians and borrowers,” explains Jaynes. “We monitor and confirm settlement of all securities and collateral movements. We mark collateral to market on a daily basis to ensure collateral is topped up to the required margins. We are tracking dealing and reporting, doing corporate action processing and dividend collection with the custodians, and substitute income payment and collection between the borrowers and the custodians. We do everything a custodial agent would do to support a lending programme in place of the custodial agent. Clients are not involved at all in the day-to-day transaction flow.”
If eSecLending takes charge of orchestrating all movements of securities and collateral, it gives lenders a degree of transparency and control impossible to obtain in a pooled lending programme. In short, it reduces operational risk. And such arrangements are now much easier to put in place than they once were. Third party lending arrangements of this kind were once highly contentious, because custodians saw themselves as servicing the needs of a competitor.
This has changed. All the major agent lenders now have sizeable third party lending franchises of their own, and the practice is common enough for custodians to set up dedicated groups to service third party lenders and their clients. “We are not dealing with the lending group, who see us as a competitor, but with the custody group, who see us as a client of their client,” says Jaynes. “We pay the custodian transaction fees to deliver assets to borrowers against receipt of collateral.”
There are limits to the reduction in risk. Obviously, lent assets cannot be segregated once they are lent to a broker-dealer (and it usually is a broker-dealer) because title is transferred. However, the countervailing collateral is held by the custodian in the segregated account in the name of the client. If it is cash collateral, its reinvestment can be managed by eSecLending, if the client chooses. However, most clients of eSecLending manage the cash themselves, or appoint a dedicated money manager.
“If the collateral is managed by another asset manager, that asset manager opens an account in the name of the client, but the cash is still delivered on our instructions from the borrower into an account in the client’s name at their custodian,” explains Jaynes. “The manager invests the cash just as it would for any other client. If clients direct the cash into a money market fund, on the other hand, it would not be segregated at all. Most of our clients prefer a segregated cash collateral management mandate.”
In fact, eSecLending clients have even secured segregated tri-party accounts. In a normal tri-party arrangement, the collateral of all lenders to a particular counterparty is held in a single account by the tri-party agent. Clients of eSecLending have not only secured a segregated tri-party account, but increasingly look to open separate accounts for each fund they manage. This is because each fund has to be treated as a separate entity and collateral cannot be pooled across multiple funds, however efficient that might otherwise be.
“Lending agents can of course show fund complexes equivalent segregation in their books, but that does not mean the collateral is held in a physically separate account for each sub-fund,” explains Lee. “So there is greater and greater call for segregation at the sub-fund level, which we can effect. If we are acting for a fund complex with multiple sub-funds, we will open a collateral account per sub-fund, per borrower. A lender with 30 funds, each with exposure to 10 borrowers, will have 300 individual collateral accounts at the custodian. This gives lenders much more control and a higher degree of transparency over how their collateral is held, and eliminates cross-subsidisation across accounts.”
In other words, eSecLending reckons that its principal competitive advantage remains performance. Pooled lending programmes, in its view, are run to optimise the returns of the pool as a whole, not its individual members. It follows that lenders that want to customise their lending programmes – say, exclude certain counterparties, or limit eligible collateral, or set higher or lower haircuts – will struggle to profit from the standardisation of terms implied by a pool, and lose opportunities to lend.
“In a pooled programme, the lender who has a collateral requirement of 108 per cent will be at a permanent disadvantage to the lender that has a requirement of 105 per cent,” adds Lee. “In our model, by contrast, both will perform on their own merits – namely, on the basis of the assets they hold and the margin they require. To us, operationally, they look the same.”
There is one regulation which is even helping eSecLending with performance. This is capital allocations. Thanks to Basel III and the Dodd Frank Act, custodians are now facing an explicit capital cost, calculated via Risk Weighted Asset formulae, to provide indemnifications to their stock lending clients. This is damaging performance, as agent lenders look to charge higher transaction fees, or take more of the lending revenue, or charge differential prices for transactions that consume capital in indemnities. Reducing the value of indemnification, or ceasing to provide it altogether, is another option under consideration.
Working out the impact of capital requirements on a particular agent lender is not easy. Capital requirements impact the universal banks, which have merged their custody arms with their investment banks, in a ways that are different from the impact on stand-alone custodian banks. Different agent lenders also allocate risk and opportunity within a pool in different ways.
In addition, although borrowers generally prefer to pledge equities as collateral, the three main forms of collateral that lenders are willing to accept – namely, cash, bonds and equities - attract differing capital charges. This means even the type of collateral an agent lender accepts on behalf of its clients will affect the charge it faces.
All of these factors increase the capital cost of providing indemnities, and the increases can be substantial. In a 2013 analysis of the value and cost of indemnifying securities lending clients against borrower default, State Street Global Markets admitted that indemnities were not traditionally factored into the price of securities lending programmes but were instead seen as “a cost of winning business.”
The paper estimated that the capital cost of indemnifying securities lending clients against loss of assets borrowed by a broker-dealer was 45 times the value of the indemnity to the client measured using an orthodox option methodology. Unsurprisingly, the capital cost exceeded the net spread on many securities loans. “Indemnification of many easier-borrow, or lower spread, trades may be harder to justify economically and perhaps ultimately may be cost prohibitive,” concluded the analysis.
“There is obviously a risk that a trade is capital–efficient for the bank, but economically sub-optimal for the client,” says Jaynes. “Equally, trades that work well for the client may not take place because the capital charge makes them impossible for the bank to support.” Agent lenders are already unwinding long term trades which are profitable for clients, but which the banks deem insupportable from a capital point of view. “The agents are saying they are not necessarily going to trade at the highest revenue opportunities for clients because they have to take into account the capital cost,” adds Jaynes. “This is a very big change in the industry that clients are starting to see, as programme revenues decline.”
As a non-depository trust company eSecLending is not impacted by the regulatory changes and capital requirements in the same way as the custody banks. The indemnities the firm offers against the loss of client assets are supported by third party insurance policies, and there is no equivalent pressure to increase the price clients pay for that insurance. “We do not have to consider different capital costs for different collateral types,” says Lee. “We are neutral on the forms of collateral borrowers wish to pledge, within customer-approved collateral schedules.”
In principle, this flexibility on collateral ought to make the supply of securities from eSecLending clients more attractive to borrowers. It would be surprising if, in this more awkward environment, borrowers were not paying more attention to their sources of supply. If the same security can be sourced from a lender unaffected by capital requirements, its price will almost certainly be lower, and its continued supply more reliable.
This is where the greater transparency afforded by the segregated approach adopted by eSecLending - indeed, the fact that its model creates an explicit relationship between an individual borrower and an individual lender, instead of delivering securities out of an anonymous pool - can help borrowers as well as lenders. “Pooled structures can be suboptimal for borrowers as well as for lenders,” concludes Lee.
 State Street Global Markets, The Value and Cost of Borrower Default Indemnification, November 2013.