The Federal Reserve published today an article called “Securities Loans Collateralized by Cash: Reinvestment Risk, Run Risk, and Incentive Issues,” by Frank Keane. On first read, the article seems to re-hash difficulties that the securities lending industry would like to forget – losses in cash collateral management accounts – without important supporting data on current trends. But the second read shows some points worth discussing.
Here is the important point:
1) The article says that even one big player in the securities lending market with overly aggressive securities lending cash collateral could cause a ripple effect, destabilizing other market actors by contagion. AIG is the poster child for this. We’ve seen it happen before and it could conceivably happen again, hence is worth keeping an eye on.
Also good but less immediately vital is:
2) “An increase in data transparency, in particular around cash reinvestment choices, seems likely to lower the possibility of runs and is a reasonable cost to bear if it mitigates the risk of financial system disruption.” We agree that transparency is helpful to financial markets in general and securities lending in particular. We are advocates, personally and professionally, of helping beneficial owners understand what is in their collateral portfolios.
Here’s what we disagree with:
3) The article implies that agent lenders are managing cash aggressively: “The range of investment strategies is determined by the agent lenders’ risk managers…. Historically, however, the disclosure practices of agent lenders and the risk limits imposed by beneficial owners have been uneven, sometimes resulting in large unexpected losses for beneficial owners and eventual litigation.” And that agent lenders have incentives to be aggressive with client money. We think that the article is focusing too much on five years ago and not enough on the present. While cash collateral losses have happened in the past, we find 0% of US pension plans or US asset managers in our repeated industry surveys that currently have cash collateral managed by agent lenders with anything more aggressive than 2a-7 fund or similar strategies.
4) The author notes the imbalance of agent lenders earning money on the upside and not sharing risk on the downside in securities lending, and cites the 2010 New York Times article on lending as an example. The author also suggests that beneficial owners find other compensation arrangements for their agents. We note that our repeated surveys show that attempts to find other solutions, for example charging a basis point fee for assets under management or on loan, have proven generally unsatisfactory. We know of some alternative arrangements but generally speaking, participants in securities lending have stuck with the current risk/reward set up. The difference is that beneficial owners are now more aware of what their collateral is invested in.
5) “An important consequence of making the main purpose of a securities loan the generation of cash in support of an aggressive reinvestment strategy is that the role of cash as collateral may be compromised. In other words, using the cash as a source
of financial leverage means it no longer serves a purely risk-mitigation function.” Yes, this is true. But regardless of the riskiness of cash collateral, investors are paid to take risk; that’s the basic premise of investing. It is the investor’s job to manage those risks and take the rewards as they come. That means supervising their cash collateral, whether in-house or run by an agent lender, in whatever format they see best. Securities lending is an investment product and some level of risk is part of the process. At this time the level of risk in cash collateral pools is very low, but still existent. At other times it might be higher. So long as investors know what they are doing (and this is a policy point mentioned in the article – see Box 2), this is what investing is all about.
Here’s what we are neutral on but want to point out:
6) The article misses some important considerations how on intrinsic value vs. collateral return is shaped. For example, Markit data show cash collateral as an important, if not critical, factor in US government bond loan revenues. We’ve seen case studies where the portion of revenue generated by collateral reinvestment for government loans was even much higher than 100%. But the article misses the conversation about the Fed’s own securities lending program that keeps Treasury GC loans at 5 bps or thereabouts, and ignores the impact of quantitative easing.
In sum, we lean towards being disappointed with this article. It’s factually fine, but the emphasis on aggressive cash collateral returns by agent lenders doesn’t make sense in the current context of securities lending. Further, the emphasis on risky agent lender collateral pools, because the agent can make more money, just isn’t accurate in our experience in today’s environment. We wish that these points had been highlighted. Since they weren’t, we think that there will be more conversation about aggressive cash collateral pools than are warranted by the reality.