Considering CVA from a Corporate perspective – implications for comparing derivatives pricing

In the first article in the series, we explained how the financial crisis of 2008 led to the rise of the Credit Valuation Adjustment (CVA) as an important factor in derivatives pricing. As CVA is an important measure of the counterparty risk in such contracts, we outlined why execution rules focussing on “best screen rate” on multidealer platforms might not result in selecting the transaction with the highest “true” fair value.

In the second article in the series, we delved further into the calculation of CVA and identified Probability of Default (PD), Loss Given Default (LGD) and Expected Positive Exposure (EPE) as key components of CVA calculation. We saw that PD increases with both the tenor of trade and the Credit Default Swap (CDS) spread of the counterparty, while EPE is determined by, amongst other factors, the tenor of the trade as well as the trade type.

In this final article we seek to quantify the influence of tenor, trade type and counterparty CDS spreads on CVA for corporates facing their banking counterparties.

Impact of CDS spreads

The CDS spread of a counterparty influences the calculation of CVA to the extent that it is used to determine the PD of the counterparty. As the CDS spread increases, the implied PD of the counterparty will increase ceterus paribus. The CDS spreads of major bank counterparties on the Bloomberg BANK screen ranged from 25bp to 144bp as of 24th June 2019 – but how does this translate into fair market value impact through CVA? For this we consider two hypothetical counterparties that approximately represent the CDS range: one with a flat CDS spread of 25bp, Bank A, and one with a flat CDS spread of 150bp, Bank B.

As a first example we evaluate a 5 year, GBP pay fixed interest rate swap. As Bank A has a lower CDS spread, this would imply a lower PD than Bank B, and therefore Bank A would show a lower CVA. In fact the difference in trade valuation between these two counterparties is around 0.044 per cent of notional as shown in Figure 1. For a £100m notional trade, this equates to the fair value of the same derivative (with the same fixed rate) traded with Bank B being £44,000 lower than if traded with Bank A. This is equivalent to a fixed rate difference of around 0.9bp annually - in other words Bank B’s price would have to be better than Bank A for the trade to have the same value to the corporate.

Figure 1: CVA as a percentage of trade notional for a 5-year, GBP pay fixed interest rate swap for a range of different counterparty CDS levels.

While this might be within the range of quotes received, as a second example we evaluate a 5-year, float-float pay USD receive GBP cross currency swap. As explained in the previous article, cross currency swaps have a more punitive exposure profile than interest rate swaps due to the final exchange of principal at maturity. For this example cross currency swap, the difference in trade valuation between these two counterparties is around 0.395 per cent of notional as shown in Figure 2. Again considering a £100m notional trade, this equates to the fair value of the same derivative (with the same quoted spread over USD LIBOR) traded with Bank B being £395,000 worse than if traded with Bank A. This is equivalent to paying an additional spread on the USD leg of around 8.1bp – in other words, Bank B would need to pay the corporate an additional 8.1bp (annualised) on each payment in order to offer a trade with the same value as Bank A.

Figure 2: CVA as a percentage of trade notional for a 5-year, float-float pay USD receive GBP, cross currency swap for a range of different counterparty CDS levels.

Impact of tenor

Trade tenor drives the CVA calculation in several ways, primarily through the EPE profile and the PD. Longer tenor trades give rise to potentially larger derivative market values, and so the cost of insuring against default increases. Additionally, the PD increases as the tenor of the trade increases. Increasing tenor, through these factors, therefore increases CVA.

Revisiting the example trades in the previous section, we start with the GBP interest rate swap. Figure 3 shows how the CVA increases with tenor of trade for both the lower and higher creditworthiness counterparties. Focussing on Bank B, with CDS spread 150bp, we notice that the impact of tenor is not linear for an interest rate swap. In fact, doubling the tenor from 1 year to 2 years increases the CVA by a factor of 7.7x, while tripling increases the CVA by a factor of 23.5x. Indeed, this relationship holds for both Bank A and Bank B.

Figure 3: CVA as a percentage of trade notional for different tenors of GBP pay fixed interest rate swaps for two different counterparty CDS levels.

Now consider the pay USD, receive GBP cross currency swap from the previous section. Focussing again on Bank B, we notice that the impact of tenor is not linear. For example, doubling the tenor from 1 year to 2 years increases the CVA by a factor of 3.0x, while tripling increases the CVA by a factor of 5.7x. Consistent factors also hold for Bank A, shown in Figure 4.

Figure 4: CVA as a percentage of trade notional for different tenors of float-float pay USD receive GBP, cross currency swaps for two different counterparty CDS levels.

Accounting implications under IFRS 13

We have so far shown that CVA has an impact on the fair market value of derivatives – but how and when is CVA recognised in the financial statements?

Under IFRS 13, the difference between the price paid for the trade and the exit price of the trade should be reflected as a loss in the income statement. CVA, where material, should form part of this difference. Whilst some corporates might currently not have to include this adjustment, arguing that the impact is not material, this argument does not necessarily hold for all exposures, particularly those of longer tenor. Therefore, should the CVA have a material impact on the trade valuation this could lead to a loss shown in the income statement, and a derivative liability recorded on the balance sheet at inception even for a trade that was traded at mid-market rates. Further, this needs to be assessed over the life of the trade, rather than at inception only. The implication here is that credit quality improvement or deterioration could lead to P&L volatility in the income statement as the CVA would change. Therefore it is beneficial not only to seek counterparties with low CDS spreads, but also those with limited volatility of CDS spreads. An additional concern under IFRS 9 hedge accounting rules could be the ineffectiveness introduced through CVA, and in extreme scenarios the ineligibility of hedge accounting should the credit risk on the derivative dominate the market value fluctuations.

Conclusions

As shown, especially for longer-dated transactions, CVA can significantly impact the “true” fair value of derivatives. Indeed, where this impact is material, there could be an accounting impact under IFRS 13 rules. In the examples shown, a £100m cross currency swap traded with two different banks, one at each end of the CDS spreads shown on the Bloomberg BANK page, could have a significant fair value difference at inception of approximately £395,000. This is equivalent to paying around an additional 8.1bp running on the USD leg of the cross currency swap – something that is striking given that a corporate policy that allocates trades based on price alone would not factor this additional cost into the decision. It is therefore vital that corporates consider the impact of CVA – not just in terms of the potential day 1 income statement loss, but also in order to avoid the selection bias that might otherwise be introduced by concentrating derivatives exposure towards the less creditworthy banking group counterparties who might only appear to offer better value.

*Large international banks refers to the group of banks selected by Bloomberg for the BANK CDS page.