Many pension funds canned their securities lending programmes outright in 2008 but have now recovered their nerve. However, Iain Morse argues that regulation is now likely to drive the future shape of the industry
Does securities lending deserve its tarnished reputation? All evidence speaks to the contrary, even though a small number of cases involving loss on cash collateral have surfaced post credit crunch. Unless trustees and pension boards are conspiring to hide the truth, it looks very much as if the complex system of contractual relationships formed around securities lending did pretty much as promised. "The Lehman unwind worked very well," notes Rob Coxon, head of international lending at BNY Mellon. "Credit has not been given to the industry for how well it worked during this period."
Despite this, lending is still seen as a little risqué, with regulators and politicians questioning whether the risks involved have been fully understood. Indeed, numerous pension funds pulled out of securities lending even before the financial crash of autumn 2008 and the international temporary bans on short selling. More than two years on, industry participants now report unprecedented levels of scrutiny of the exact terms of conditions of business between lenders and custodians. "Trustees must be happy with the risks involved; there is a sense that these were not fully communicated in the past," warns Hugh Gittins, a senior associate and pension law specialist at law firm Eversheds.
Collateral is the key to lending risk. Put up by the borrower, this can take the form of cash, equities, fixed income, or a combination. If the borrower defaults, the collateral compensates the lender. The value of collateral will always exceed the value of borrowed securities in ratios which, although not set by law or regulation, are widely used. These typically range from 102% to 105%. The excess is called the lending ‘margin', with higher margins required where cross-currency borrowing takes place with attendant currency risk.
"There is a demand from pension funds to customise and risk adjust all aspects of their lending and this includes more attention to collateral," notes Paul Wilson, managing director at JP Morgan World Wide Security Services. By tradition, the US, whose banks were the first to offer securities lending, use cash collateral, while Europe tends to prefer equity or fixed income. But this picture is changing rapidly. According to Data Explorers, which provides industry data, the global use of cash collateral as the source of lending revenue stood at 67% of all collateral in January 2007, falling to 59% and 57% over the following two years, but shooting up to 79% by January 2011.
Cash collateral revenue in the US has declined slightly - from 96% in 2007, to 95% in both 2009 and 2010, and now standing at 91%. In the UK, it stood at 18% in 2007, 15% and 21% in 2009 and 2010, respectively, and is now increasing to 38%.
In Europe, matters have been different, with respective figures of 35%, 32%, 29% and for January 2011, only 23%.
Over the same four-year period, revenues in Australia fell from 75% to 66%, but rose in Canada from 37% to 56%. Whatever doubts might have pervaded the market, cash collateral looks to be back in fashion.
The unwary might ask what risk of loss there could be in holding cash collateral. The answer is that pre-credit crunch, cash collateral was split into risk and duration-adjusted portfolios to increase returns. Cash collateral on deposit earns interest and, to begin with, collateral pools might be split into overnight, one-month and six-month deposit accounts to gain a few basis points on interest. The largest volume of lending is overnight, so a duration mismatch opened between lending and collateral durations. But as lenders usually continue lending as part of a ongoing programme, this mismatch could be calibrated to a virtually risk-free level.
However, as the credit bubble extended, the composition of cash collateral pools gradually became more heterdox. For example, a pre-crunch cash collateral portfolio might have 40% exposure to overnight lending, 20% to one-week, 20% to one-month and 20% to six or even 12-month lending. If this portfolio were pooled, as most were and are, past experience of withdrawals by programme lenders and consequent repayment of collaterals will have been factored into this mix of portfolio durations. Only if there are exceptional levels of withdrawal will longer duration deposits in the pool have to be liquidated, with consequent loss of interest penalties.
This prospect might not seem too potentially damaging. The problem comes, however, if the longer duration part of the portfolio is used to purchase short-term portfolios of consumer debt and not held in cash deposits; portfolios of residential mortgages, credit card debt and car finance debt for instance. Default rates will have been factored into these portfolios to provide the purchaser with greater security, but these default rates are data-mined from past averages. In the credit crunch and ensuing collapse of sub-prime mortgages in the US, past experience proved no guide to the future. Default rates rose to unprecedented levels.
This, in turn, rendered some of these longer dated, asset and non-asset backed portfolio components illiquid. If they could not be sold, or in some cases sold only at a huge capital loss, the cash collateral could not be recovered in full or paid to the lender in case of counterparty default. A couple of banks defaulted, while others collapsed to knee height before ‘rescue', while the US sub-prime housing market collapsed.
The mathematics on how this might affect a collateral pool need closer examination. Assume cash collateral of €102m or 102% of the amount borrowed. Suppose 20% of this portfolio is in longer consumer debt. This debt has been purchased with a built-in default rate of 10%, but this rate unexpectedly doubles to 20%. As a result, the €20.4m so invested is worth only €18.36m, leaving the total collateral with a value of only €98.36m. Ouch! Trustees might have thought that cash collateral was risk free and now discovered the contrary. In reality, collateral pools were much larger than this, with aggregate values running into billions of dollars. But such losses did take place.
There is an alternative to accepting losses on collateral. Lenders can stay in-programme, stay lending in other words, and therefore balance sheet losses on collateral are not realised. This, effectively, is what happened to some lenders that might have wanted to withdraw or trim their lending programmes post credit crunch. This is a situation that trustees and boards will want to avoid in future, if some even more unexpected and extreme credit event were to take place. "Durations on cash collateral pools have come in since the crunch," adds Coxon - the credit ratings on fixed interest used as collateral have been raised, the margins of equity collateral increased.
A frisson of anxiety ran through the pensions industry in 2008 regarding the risks that a pension fund is exposed to when engaged in a lending programme. Part of the problem lies in the pre credit crunch perception of lending. It was not well understood, hardly discussed in detail by trustees and not seen as a source of uncompensated risk. The custody banks acting for pension funds and many other institutions with stock to lend liked things this way. On thin margins, the larger the programmes and collateral pools, the more revenue they received. Lending, then, helped to pay for the ever increasing range of non-core services offered by the custodians to their clients, particularly pension funds.
All the evidence suggests that institutional lenders, including pension funds, have recovered their nerve and are keen to lend. Reducing risk in collateral pools has been accompanied by far greater attention to counterparty risk; some lenders will only lend to counterparties with a very high credit rating. There is also more attention to setting limits on the percentage of portfolios that can be lent, with some lenders reducing this amount.
The market as a whole has shrunk, but not because of reluctance to lend. "There has been a contraction in borrowing," warns Wilson. "The demand side has been impacted by shrinkage in the hedge fund sector, too much supply of fixed income, and equity traders going long."
There are regulatory impediments to the lending industry ahead. If banks are required to offer indemnities, then they will need to increase their capital ratios against any liability to stock lenders. "This will put up the costs of the industry," says Coxon. Some banks, under balance sheet pressure from multiple sources, may decide to leave the industry.Then there is the prospect of central counterparty (CCP) platform. "A CCP does not need to offer indemnities," Coxon adds "They charge margins, and would disintermediate parts of the lending industry." We might also see new restrictions on the hedge fund industry. In the end, EU politicians may reshape and reduce the size of the industry.