Despite a focus on liquidity since the global financial crisis, the supervision and regulation of liquidity still lags far behind that of capital, say three central bankers with De Nederlandsche Bank (DNB).
Iman van Lelyveld, Paul Hilbers and Clemens Bonner are the editors of Liquidity risk management and supervision, published by Risk Books in December 2015. In an interview with Risk.net, they say the focus on liquidity is the result of hard lessons learned during and after the crisis.
"There were already liquidity regulations in some countries in the 1980s and 1990s," says Bonner, a policy adviser at DNB. "The Netherlands was one; the UK and Germany were others. The Basel Committee [on Banking Supervision] had already started to talk about harmonising liquidity regulations in the 1970s. But that never really did get traction."
The main problem, says Bonner, was an inability to understand and model how liquidity regulations would interact with monetary policy decisions.
Van Lelyveld, a senior policy adviser at DNB and professor of banking and financial markets at the Free University of Amsterdam, says some simply did not see liquidity as a valid regulatory concern. "Markets were expected to focus just on solvency and the capitalisation of a bank when there was an issue of funding," he says.
However, the 2008 crisis showed it was possible for even well-capitalised banks to run into liquidity shortages. During the crisis, liquidity stresses were a key factor in the downfall of banks such as Lehman Brothers, Northern Rock and Washington Mutual. Consequently, liquidity risk management has climbed up the agenda of banks and regulators alike – a trend that senior supervisors expect to continue in 2016.
In a foreword to the book, Stefan Ingves, governor of the Swedish Riksbank and chairman of the Basel Committee, writes: "The global financial crisis reminded us of the need for sound liquidity risk management. Strong liquidity risk management – in both its quantitative and qualitative dimensions – is undoubtedly critical to long-run success in banking."
Basel III, finalised in 2010 and revised in 2013, sought to address the issue of liquidity more thoroughly than Basel II. The rules include two new liquidity ratios banks must meet: a liquidity coverage ratio (LCR), designed to ensure banks have enough high-quality liquid assets to meet a one-month period of severe market stress, and a net stable funding ratio.
But regulators are still getting to grips with the idea of supervising liquidity, says Bonner – particularly when it comes to stress testing. "Even on the capital side, there is still a long way to go before we have really advanced capital stress tests, but of course we are a lot further there than we are on the liquidity side."
Liquidity stress-testing practices still differ a great deal from one country to the next. More advanced central banks, such as the Reserve Bank of Australia, run individual bank tests and market-wide tests using a wide range of scenarios, notes Hilbers, a director of financial stability at DNB and member of the Basel Committee. Some don't perform any liquidity stress tests – relying only on the LCR as "a sort of short-term stress test", he says. Other countries run limited tests for individual banks, or use only a small number of scenarios.
One problem is that liquidity stress tests are inherently trickier compared with those for capital, says Bonner. Many banks are realising this as they adopt the LCR – a requirement that is supposed to be fully phased in by 2019. "When a bank breaks it is relatively easy to say 'this is the amount of capital they would have needed in order to survive'," says Bonner. "With liquidity, that's much more difficult, because you don't know how long the runs would have continued if the bank had survived."
If the bank comes with a plan that says 'in that case I will go to the central bank', that is a big no-no
Iman van Lelyveld, DNB
In general, Bonner believes individual banks should be running liquidity stress tests using both bank-specific stress scenarios and market-wide stress scenarios. While a market-wide stress scenario may be less severe in terms of liquidity outflows, having all banks under pressure at the same time would make it much harder for an individual bank to obtain funding, he says. But many banks have struggled to perform such market-wide tests – in part, due to a lack of information about best practice.
Market-wide tests can also give rise to thorny modelling challenges. In a November 2015 working paper co-authored by van Lelyveld, the Basel Committee emphasised the need for market-wide tests that included the impact of liquidity in their models. This means they must consider what van Lelyveld describes as the "feedback effect" of liquidity – for example, by simulating the funding shortfall affecting profit and loss, knock-on effects on capital, and how these in turn can change funding risks and costs.
When supervising liquidity stress tests, supervisors must scrutinise carefully the extent to which firms assume they can rely on central banks as lenders of last resort. Bonner portrays this is as a balancing act. On one hand, banks should not necessarily be prevented from assuming they will retain access to central bank liquidity. But on the other, regulators must enforce the LCR's stipulation that banks should not seek central bank help immediately, but instead internalise the cost of taking liquidity risk.
"In the first phase – be that a month or six weeks – you wouldn't expect banks to have to rely on the central bank," says van Lelyveld. "There is always ambiguity there, because you don't know what the central bank is going to do. If the bank comes with a plan that says 'in that case I will go to the central bank', that is a big no-no."
The best way to tackle the issue is through reverse stress testing, he adds. If banks are able to show they could survive without central bank funding for 10–12 weeks, regulators should feel relatively comfortable. But if the stress scenario led to the bank seeking central bank funding within a few days, "you might get concerned".