Regulators cannot find the liquidity problem that traders see daily, threatening to set rules for game ‘A’, when game ‘B’ is played out in the markets. Chris Hall reports.
Regulators and market participants are at loggerheads over the liquidity problem. Defining liquidity, more than any other characteristic of the financial markets, depends on context. Broadly speaking, a liquid market is one in which buyers and sellers can find each other reasonably quickly and cheaply, and transact at a mutually acceptable size and price. Finding common ground on precise metrics is harder. Yet regulatory reform will mean these definitions and metrics could influence future market structure.
The European Commission recently recommended to the European Securities and Markets Authority (ESMA) that MiFID II’s new transparency rules for the European bond markets be reduced in scope, phased in and subject to annual recalibration, to minimise disruption to liquidity. If the Commission endorses ESMA’s revisions, buy-side market participants will heave a sigh of relief. The subsequent changes will grant them vital extra breathing space to adjust their investment strategies and operations to accommodate the bond markets’ already evident shift away from a liquidity-sourcing model based on principal-based market-making by broker-dealers.
However the regulatory perspective on liquidity clashes with that of traders. Anecdotally, buy-side trading desks report increasing difficulties finding the issues they want in the sizes requested by portfolio managers. They cite wider bid-ask spreads, falling trade sizes and – admittedly imperfect – transaction cost analysis reports which show costs are rising as orders take longer to fill.
In February, however, the UK’s Financial Conduct Authority (FCA) published a research paper entitled ‘Liquidity in the UK corporate bond market: evidence from trade data’, which used academically rigorous metrics to argue that liquidity levels in the UK corporate bond market had not deteriorated significantly between 2008 and 2014. If regulators adhere to a measurement model which is flawed, it could lead to rules or benchmarks being set that fail to account for actual market circumstances.
Market participants have no quarrel with at least one observation made in the FCA paper. The research asserts that sell-side inventory levels have declined by almost half in little over five years: from £400 billion of UK corporate bonds held by primary dealers in mid-2008 to around £250 billion toward the end of 2014. A similar trend has been identified in the US and other major markets, representing an historic migration of assets to buy-side balance sheets, due in part to higher capital costs for under Basel III.
Despite this, the FCA report characterises the UK corporate bond market as remarkably resilient since recovering from the initial shock of the global financial crisis. Far from being squashed by Basel, “the market appears to have become more liquid in recent years”.
The report does not deny that periods of market stress can lead to a deterioration of liquidity, acknowledging a marked fall-off in liquidity levels as market stresses heightened in 2009/10. However, it concludes regulatory reforms have not caused reductions in liquidity in normal market circumstances nor caused liquidity to become more ‘flighty’ in abnormally stressful market conditions. Even during the taper tantrum of Q2-Q3 2013 or the flash rise of Q4 2014, “we cannot see any meaningful spikes in the liquidity risk measures”.
Nor was the UK somehow isolated or shielded. The FCA’s analysts point to parallel trends in the French and US bond markets, as reported by local regulatory bodies. In addition, the report finds no evidence that liquidity now accounts for a greater proportion of yield spreads on UK corporate bonds than previously, which would have suggested higher cost of capital for issuers arising from worsening liquidity conditions.
In terms of methodology, the researchers eschewed market measures of liquidity. Noting the limitations of quoted bid-ask spreads (“quotes are usually non-binding and there is no minimum quantity attached to the quoted prices”), they selected five measures commonly used in academic circles. The report draws on data for all trades in corporate bonds for which the FCA is the national competent authority, including price, quantity, time and counterparties (albeit cleansed to eliminate outliers). But it is based on two vastly different sample sizes. The reporting system used by the FCA up to August 2011 (Sabre II) only yielded data on 406 bonds, while its successor (Zen) provided the researchers, who acknowledge the scope for variation between samples, with data on 6,291 UK corporate bonds.
Aghast that the FCA should produce statistical evidence that flies in the face of everyday trading experience, market participants have thrown various barbs – targeting the data set rather than the methodology. The data set is incomplete, not transparent, biased toward highly liquid instruments, or covers the wrong time period, according to detractors.
David Krein, head of research at MarketAxess, observes that the FCA sample lumps sterling-, euro- and US dollar-denominated bonds together (39% of the larger ‘Zen’ sample are euro-denominated), in contrast to market practice, which treats bonds issued in different currencies as separate markets with distinct dynamics. Krein’s analysis of trades in investment grade names reported to MarketAxess’s TRAX post-trade utility shows bid-ask spreads were wider and more variable (between 50-80 bps) for sterling-denominated investment grade bonds versus euro-denominated equivalents in the period from July 2013 to March 2016. “The fact that the FCA report mixes sterling- and euro-denominated bonds diminishes its usefulness,” he explains.
But the biggest problem with the data is that it does not tell you what people could not trade, but wanted to. This might seem obvious, but it is particularly important in the corporate bond market, where the vast majority of ISINs are terminally inactive six weeks or so after issuance. Unlike equities, where traders can always get filled if they’re prepared to pay, bond traders who cannot find liquidity in size for the issue they want must try to find a proxy from the same country of origin or with a similar capital structure.
As such the buy-side traders’ increasing challenges in pursuit of liquidity may not show up in the metrics, because they manifest themselves in ways not captured by available data. But they are no less tangible for that. “At present, there is very little depth to the market, and very little certainty on whether a trader will in the future be able to buy back at a fair price a bond that he sells today,” says Lee Sanders, head of FX and fixed income trading at AXA Investment Managers. “Banks are reluctant to put balance sheet against smaller trades. When traders are struggling to find a price for a E500,000 trade in an investment grade name, that suggests there is a problem with liquidity.”
Acting on the evidence presented by their trading screens, market participants are responding to the aforementioned shift in corporate bond inventory. “When looking for a price, buy-side firms would historically go to a small number of sell-side counterparts. But because these firms are offering prices in fewer names and smaller sizes, the buy-side are having to reach out to more counterparts. Liquidity is there, but it’s not necessarily being generated by brokers,” says Mark Russell, global head of clearing & execution services, cash products at UBS, who argues that the automation of bond trading will have to take a different path to that seen in equities and FX to account for the market’s inherent illiquidity.
At MarketAxess, the need for buy-side firms to cast a wider net led the platform to first abandon its cap on RFQ recipients, and then launch all-to-all trading in 2014. The firm recently reported 82,000 transactions completed via its Open Trading protocol in Q1 2016, more than double its Q1 2015 level.
AXA IM’s Sanders is optimistic that technology will help the buy-side to conduct a more efficient, yet more diverse liquidity search. “We’re continuously working to have an aggregated pre-trade view of liquidity, and then pulling in firm, actionable prices from multiple sources into our EMS is a real focus,” he says. Sanders predicts greater execution choice for buy-side firms, deploying liquidity matching engines, smart order routers and open protocols to source liquidity, or working with liquidity providers on ‘large-in-scale’ orders. “Ultimately we’re all governed by best execution obligations,” he notes.
MiFID II’s transparency rules are not the only upcoming threat to liquidity. Andy Hill, senior director, market practice and regulatory policy at the International Capital Markets Association, says the Central Securities Depository Regulation’s planned mandatory buy-in regime will be “highly detrimental to market-making in the European corporate bond market”. Nevertheless, the European Commission’s decision on MiFID II may prove crucial. A phasing in of transparency rules may give bond market participants precious time to put the necessary systems and connections in place to tap liquidity effectively.
In a market crying out for reliable data, transparency is still regarded as a double-edged sword. The bond markets’ structural shortcomings in data generation thwarts decision-making, by regulators and participants alike, but, as the gradual evolution of the US post-trade TRACE feed suggests, change takes time. “Aggregated, market-wide pre-trade data would help us understand what is really going on in the market,” notes Sassan Danesh, managing partner, Etrading Software, and Project Neptune Management Team. “The market has to come up with a solution that is appropriate for the transparency requirements of the product set.”