By Alex White, Head of ALM Research, Redington
The most basic rationale for LDI is to hold instruments that pay money if schemes’ liabilities become harder to afford. Essentially, if schemes need more growth assets to pay liabilities, they need an asset that provides more money. That means they need more than just the valuation uplift; they need to be able to realise the money and re-invest it. Illiquid LDI therefore doesn’t work –it must be possible to realise cash to pay liabilities.
Swaps are technical, complex instruments that allow you to trade one set of cashflows for another. You can ask to break or unwind a swap – however this is effectively a new contract – there is no break fee and valuation baked into the original swap. The exit price is at the discretion of your counterparty. Historically, this has led to a few issues emerging.
Before the global financial crash, interest rate swaps were typically projected and discounted at 6-month Libor. This made the maths very clean as you had one interest rate and you knew the first fixing.
During and after the GFC, 1-day cash and 6-month traded differently, and swaps started to discount at SONIA.
Subsequently, the CSA (Credit Support Annex) terms started to matter. These dictate what can be posted as collateral. Some CSAs give both counterparties the ability to post other instruments, such as corporate bonds, or the cheapest to deliver of a basket of multi-currency government bonds. These are “dirty” CSAs. The potential to post instruments other than cash effectively leaves the bank long and short two options on their capital position, which is priced into a discount spread over SONIA. This led to realised valuations becoming dependent on bank-specific capital positions
Similarly, the same issues came up again with cash and gilt CSAs compared to pure cash CSAs. Gilts incur a less favourable treatment on bank capital than cash under leverage ratio rules. This means the realisable value on swaps with such CSAs may be closer to the LIBOR discounted value.
Clearing incurs margin, but does seem to circumvent the risks of similar outcomes. However, CSA issues only directly affect price on exit, which means they can have the opposite effect on entry prices. This has led to some curious aspects of the market, such as a bilateral vs cleared pricing basis. At the time of writing this has been as wide as 5bps in inflation.
Other issues are also still ongoing. As a reaction to the LIBOR-rigging scandal, LIBOR is going to lose FCA support in 2021, and it will need to be replaced with transaction-based referenced rates. In the UK, it will largely be replaced with (reformed) SONIA. In 2021, it is likely that definition of LIBOR will change, probably to SONIA plus some term structure of fixed spreads. It is hard to envisage this sort of change will not result in some clear winners and losers.
In practice, these issues do not undermine the strategic rationale for using swaps. They do have implementation implications though, and they highlight the fact that swaps do not so much eradicate repo roll risk as incur alternative risks instead.
|
