The mere mention of the term funding valuation adjustment (FVA) can trigger long, heated arguments between quants, traders, auditors and academics, mostly focused on three broad issues: whether it should exist; how it should be accounted for; and how it should be priced. In most cases, wider debates on the topic branch into a number of more specific arguments over the finer points.
To stop FVA from getting any more divisive, it is important to line up all the tools and information available, to clarify where the roots of the disagreements lie – how to define the economic value of FVA.
In this month’s first technical, Derivatives funding, netting and accounting, Christoph Burgard, head of global risk analytics at a US bank in London, and Mats Kjaer, head of quant XVA analytics at Bloomberg, throw some light on this fundamental question.
The quants present a framework for pricing FVA based on two major market views. The first, proposed originally by the quants and already widely adopted in the market, assumes funding is symmetric – that is, you can borrow and lend from the treasury at the same rate.
The other view, proposed by Claudio Albanese, Leif Andersen and Stefano Iabichino in 2015, argues that funding is asymmetric; that trading desks have to fund their derivatives at the funding spread, but if they have excess cash, they can only reinvest at the risk-free rate, which is lower than the funding spread. Excess cash in one netting set, however, can be used to fund other netting sets, which the quants show leads to a lower FVA.
Using the two views, Burgard and Kjaer extend their widely used replication framework for pricing FVA to multiple netting sets and present three funding strategies. The first two align with their own view on FVA. The third strategy uses the asymmetric funding assumptions of Albanese, Andersen and Iabichino.
The quants then calculate funding costs for the three strategies on long-short and long-long positions of forwards with two counterparties. They find the asymmetric funding strategy is the least profitable approach in terms of shareholder economic value. In the long-long case, the funding cost was almost double the first two strategies.
Burgard argues the symmetric funding assumption – which is more profitable in their example – aligns with market practice.
“In practice, what happens in a bank is the derivatives portfolio, if it were positive funding, will generate cash and it will go through the central treasury department of a bank and then will be reused somewhere else that requires funding.”
This view however, has its detractors.
“I think most banks’ treasury departments would struggle to agree to remunerate excess collateral above the risk-free rate,” says one XVA expert.
However, Burgard says one of the aims of the paper was to provide a link between funding strategy and accounting considerations.
“Accountants are driven by their general accounting principles. The question, though, is how you apply principles in this case to achieve a good alignment between what you account for, say the common equity tier 1, which is supposed to represent shareholder value, and what it is worth to them; in other words, how do you account for it so it aligns with the economic value to shareholders you can generate over the life of the trade,” says Burgard.
In their past papers, Burgard and Kjaer have presented FVA as impacting the economic value to shareholders. Current practice of deducting FVA from earnings aligns with this. Quants such as Albanese and Andersen, and academics such as Stanford’s Darrell Duffie, argue FVA is a transfer of wealth from shareholders to bondholders, which net out in fair value and therefore should be deducted from equity rather than earnings.
Some see this as a choice between shareholder value versus overall bank value.
The different views, and the research that supports them, highlight just how many stakeholders have come together to sort out FVA once and for all. Banks started pricing FVA, and so set market practice. Quants and academics have done their job of presenting what values it can take under various assumptions.
Perhaps if standard setters, by working closely with the industry, offer more clarity on which assumption is the most practical and reflective of economic value, most of the other disagreements will drop out of the equation, one by one.
All the tools and information are on the table – it’s time to set the standards.