A new staff working paper from the Bank of England finds that clients trading interest rate swaps (IRS) in an uncleared environment are paying around eight basis points for the privilege.
The paper uses data from trade repositories to study trading and pricing patterns in IRS markets and finds the risk premia attached to, and therefore the pricing of, IRS trades varies greatly. The price differentials in risk premia are, “highly significant in statistical and economic terms,” the paper states.
This premium substantially decreases when initial margin is posted and with the client’s creditworthiness, the paper adds, it then goes on to rationalise the premia with increased inventory costs induced by recent financial regulation that are passed on to market prices through so-called valuation adjustments (XVA).
The paper also notes that OTC premia favour dealers in the sense that the extra cost in the non-CCP segment for interest-rate protection (i.e., paying the fixed rate) is absent for dealers and large for clients. This suggests regulatory costs and bargaining power for dealers, the paper argues.
The authors, Gino Cenedese, Angelo Ranaldo and Michalis Vasios, observe that one of the declared objectives of the G20 post-GFC, was the increase of standardisation, greater use of central clearing, and the reduction of systemic risk, especially in terms of credit risk. “Two questions then arise,” the observe. “First, have the regulators achieved their objectives? Second, have market participants efficiently adapted their pricing and trading activities?”
At one end, the authors note that more expensive non-CCP derivatives incentivising central clearing can be regarded as a suitable achievement, however they also observe that more diversity in the form of adverse prices for dealers’ customers would be an unintended consequence.
The paper maps the IRS market in the new regulatory regime, which represents the largest part of the entire interest rate derivatives market by analysing every contract by non-US-based counterparties (US participants were subject to mandatory clearing at this time) that traded with UK-based entities (including subsidiaries of foreign banks) and that was reported to the Depository Trust & Clearing Corporation (DTCC) between the beginning of December 2014 and the end of February 2016. The time span intentionally starts after the US clearing mandate, introduced in March 2013, but before the European one in June 2016, the authors note, explaining this means that the counterparties in the sample (with access to CCP) had the choice whether to centrally clear their trades or not throughout our sample period.
A unique feature of the data, the authors argue, is the information on counterparties’ identity, which allows them to identify and quantify a trader’s credit risk, its degree of financial sophistication, and those of its counterparties.
The authors claim that “clear results” emerge from their study. First, they say they uncover new stylised facts of the IRS market in the new regulatory regime. These are that cleared trades are the dominant variety in terms of market share and these trades are generally more standard contracts with larger notional and longer maturity. They also find that the average market participant on CCPs is a more “sophisticated” financial firm with higher creditworthiness consistent with the theoretical predictions that the non-CCP venue concentrates higher counterparty risk. They also suggest that in distressed markets, market participants “significantly increase” their trading outside CCPs suggesting that they tend to circumvent the so-called margin procyclicality, i.e., tighter funding conditions and higher margins imposed by the CCP.
As noted, when a client buys interest rate protection pays a dealer a fixed rate that is around 8 basis points higher for a non-CCP transaction relative to a CCP one, something the authors argue is “a sizeable amount” given the total notional traded of around $7.4 trillion and compared to the average swap rate over the sample period. The PTC premium reduces by at least half if a non-CCP trade involves posting of initial margin, however, and through the lens of XVA, the results point to credit risk (CVA) and capital charges (KVA) as important drivers of OTC premia for non-CCP transactions.
The paper also finds that the premium is not fully symmetric when dealers receive or pay the fixed to their clients. It is “significantly positive” when dealers receive the fixed rate (consistent with dealers passing on XVA ‘costs’ to clients) but not systematically negative (i.e., there is no discount) when dealers pay fixed.
“More precisely, when dealers pay fixed to clients (irrespective of clients’ type), OTC premia tend to be smaller but still positive,” the authors write. “When we restrict the sample of clients to non-banks only (so we remove transactions by non-dealer banks), OTC premia tend to be negative.
“Our tentative explanation is bargaining or market power: dealers pass the XVA cost to non-banks, but they are less able to do so against other banks who are also subject to the XVA costs,” the continue. “In addition, it is important to stress that in our sample period the expectation for a rate decrease was virtually non-existent, as suggested by the steepening OIS curve. That is, there was only an upward risk of rate hikes by the Fed. Thus, across our sample period the payer of fixed (including dealers) pays a premium to protect herself against the interest-rate risk. Overall, these findings are consistent with the idea that (i) as any inventory costs, dealers tend to pass on regulatory costs to market prices, and (ii) dealers exercise some bargaining power when they price swap contracts.”
The paper also identifies the main determinants of the OTC premia. Specifically, it finds that swap prices tend to increase with notional amount and remaining time to maturity, which capture the contract risk, and that agents with higher creditworthiness tend to obtain better prices. Conversely, the price a client bank pays or receives for the fixed rate to or from a dealer shifts with the bank’s credit risk consistent with the pricing of CVA and KVA. Translated into numbers, the paper says that if the bank’s credit risk is one notch below as assessed by credit-rating agencies, the swap rate is one or two basis points more expensive.
Finally, the paper analyses the role of market liquidity, relationships, and bilateral exposures in OTC premia. To do this, it uses data from Swap Execution Facilities (SEFs). It finds that IRS prices through SEFs are smaller but the OTC premia remain statistically and economically significant.
While the authors observe that their analysis does provide evidence of a shift in behaviour towards the greater use of the “safer” CCP structure, and therefore can be considered a positive outcome for the G20 objectives, they also note that evidence of price heterogeneity and dealers’ bargaining power against their clients means that transparency and efficiency, which are also declared objectives, can be improved.