The events surrounding the London interbank offered rate (LIBOR) fixing in 2012 brought financial benchmarks into the public conscious for the first time and set off an important chain reaction. Within one year, LIBOR would be regulated. And soon after a number of other key benchmarks would be too, as policy makers sought to bring confidence to rate setting procedures underpinning trillions of pounds of contracts in markets ranging from gold to oil, silver and foreign exchange.
One of the biggest of those benchmarks - by underlying market volume – is the ICE swap rate. Prior to coming under the regulatory umbrella, the ISDAFIX, as it was known, was set every day at 11am and based on the mid-point of prices that a panel of 16 banks were willing to trade standard market size (SMS) swaps.
In common with other benchmarks at the time, this process was pretty much defined by old fashioned financial principles of dictum meum pactum - my word is my bond. Honourable as that sentiment was, the panel process lacked objectivity, transparency and oversight. Moreover, it invited questions about accuracy. Were the submissions of the 16 banks really representative of the fundamentals in the underlying swap market?
The ICE swap rate lies toward the opposite of the spectrum. Contemporary and data-led, it is based on continuous, randomised snapshots of order books from multiple trading venues. The benchmark now reflects what is tradable in the swap market. It is also subject to better governance and greater oversight by a diverse committee of both independent and market representatives.
So on March 31, 2015, when the unregulated ISDAFIX became the regulated ICE swap rate, policy makers expected the new process to boost confidence in the accuracy of the rate setting process.
But there was also caution. There were fears the new rules could create unintended consequences in the $289tn swap market and possibly reduce its efficiency. Recent analysis (set out fully in our occasional paper) of data collected by the FCA, however, tells a positive story.
More accurate and more liquid
Broadly speaking, data on 10-year USD contracts (the most widely traded in our sample) suggest the market is more liquid today than it was under the ISDAFIX regime.
Order book data from Trad-X – one of the UK’s largest electronic derivatives trading venues – shows the average implied execution costs for (10Y) USD contracts narrowed from 0.78 bps to 0.7 bps - a reduction of 11% at the 1% significance level. In other words, it is cheaper to trade electronically under the ICE swap rate regime.
Ultimately, this conclusion is based on a “fill spread” measure (the estimated roundtrip costs for completing a buy transaction and a sell transaction) using order book data from August 1, 2014 to December 30, 2015. Measuring liquidity accurately in this particular market is of course tricky. But the average daily time-weighted fill spread is a solid proxy.
Looking back through that order book data, two periods stand out (see figure one). The first is a spike on 4 December, 2014, where liquidity declines. Likely due to a drop in the number of participants in the US dollar contracts on the trading venue. A pretty credible scenario given the fact we saw the number of dealers on the platform drop by some 45% the next day.
The second and more important period is on March 26, 2015, just three trading days before the benchmark regime changed. Here we see a marked improvement in liquidity. The theory being that more traders entered the market as price transparency improved, stimulating stiffer dealer competition on prices and driving out the weakest. An argument based on recent scholarship by the academics Professor Darrell Duffie at Stanford and his co-authors.
Given the proximity of this improvement to the change in regime, and the fact there was no major market-moving event on switchover day itself, there seems to be a connection. This adds weight to the argument that a robust and regulated market-based benchmark reduces information asymmetry and signals integrity, encouraging greater market participation. The fact that on the 26 March 2015 the Trad-X platform experienced a 10% increase in participation, reaching an all-time high, supports that hypothesis.
Figure 1: Time-weighted fill spread
Notes: This figure shows the development of the daily average time-weighted fill spread for the 10Y USD IRS over the sample period.
Nonetheless, you can’t just infer that regulation caused the changes. You need a control group. And if you take two interest rate swap classes – one affected by the new regime, one not – and lay them side by side, statistically speaking, the contrast is telling.
The 10Y USD IRS is affected by the new ICE swap rate regime and shows a significant enhancement in liquidity after the change in regime. The 12Y USD IRS, for which no rate is assessed, shows less improvement. The results are robust to using more sophisticated specifications and controlling for confounding effects.
What does this mean
Why does this matter? On one level, there is certainly a direct benefit to dealers themselves. On the Trad-X platform, the cost savings for electronic transactions in 10Y USD interest rate swaps amount to around $7 million from April 2015 to December 2015 alone. In other words, we are seeing more efficient markets.
It is important to note though, that the benefits of the benchmark regime change are not isolated to individual firms. They have knock-on consequences for us all. The ICE swap rate is used by many firms to hedge interest rate risk and thus to reduce uncertainty. March 31, 2015 was – it appears – an important step forward.
Our economists’ full occasional paper on the impact of benchmark regime change in the interest rate swap market is available below. It is the first in a series of research papers assessing the ex-post impact of the regime on different benchmarks.
This article and its accompanying research were co-authored by Gbenga Ibikunle, Associate Professor at the University of Edinburgh and Vito Mollica, Assistant Professor at Macquarie University. Our thanks go to both.