What is causing flash crashes?

And how do we stop them?

December 30, 2016
British financial trader Navinder Singh Sarao ©Matt Dunham/AP/Press Association Images
British financial trader Navinder Singh Sarao ©Matt Dunham/AP/Press Association Images

Flash crashes in financial markets are causing increasing consternation among central bankers, traders and indeed anyone who depends on the great many things that are priced by markets. In early December, the Bank of England’s Prudential Regulatory Authority published its report on October’s flash crash in sterling. In the early hours of 7th October, the value of the pound fell 9 per cent, a slide that occurred in just 40 seconds. This was just the latest in a series of flash crashes. The first took place on 10th May 2010 when during just 36 minutes, US stocks lost $1 trillion of value before largely recovering.

Many politicians, prosecutors and journalists believe in the “Bad Apple” theory, in which these events can be traced to the actions of rogue individuals. After the first flash crash, they set off in pursuit of the evil-doers. Five years later, their chase ended at the door of a semi-detached house in Hounslow, west London where a 36-year-old man, Navinder Sarao, traded stock futures on the internet. Similarly, Bank of England investigators believe that sterling’s October flash crash was caused by a lone trader working for Citibank in Tokyo.

Sarao was extradited to the US and pleaded guilty to two of the counts he was charged with in November. But some have questioned whether he was solely responsible for the flash crash. And the scale of the allegations against single traders does not sit well with regulators' claims that, after several years of toil, they have delivered a more robust and resilient financial system.

Flash crashes are not only more frequent but they are occurring in markets where stability is both expected and critical to the smooth function of market economies. The German government bond market, which is the base for determining the interest rate payable for debt in Europe, experienced a flash crash on 15th January 2015. Outside of Europe, more financial contracts derive their price from the US government bond market than from any other, but even the Treasury market proved susceptible to a flash crash on 15th October 2014.

Why do flash crashes happen and what can we do about it? Let’s start with how financial market function. People own assets, such as a home or shares in companies. These assets are considered liquid if there is always a ready buyer prepared to pay the price close to their long-run value. Liquidity in financial markets needs diversity. It requires someone willing to buy when others are selling because they have a different view of the value of the assets, or a different time horizon, or a different purpose, or some other difference.

Buying when others are selling requires having the funds to absorb any losses in the short-term in anticipation of gains in the future. These funds, referred to as capital in the language of financial regulation, have to be yours to lose as opposed to funds you have to pay back to someone else. Consequently, savers and those who manage their savings in a conservative manner such as pension funds and life insurers, along with well-capitalised banks, are the natural suppliers of financial liquidity when markets are choppy.

However, in a notable example of unjoined up policymaking, bank regulators have acted as if insurers were better providers of liquidity than banks—while insurance regulators have been acting as if the opposite is the case. Bank regulators are making banks hold three times more capital against the risks of their trading positions going sour than before the crisis. At the same time, regulators of life insurers and pension funds are basing new capital requirements on the short-term volatility of their assets, rather than the likelihood that these assets would fail to achieve the requisite long-term return.

This rule makes it more costly for life insurers and pension funds to buy assets that have crashed lower, even though that is exactly the right time for long-term investors to do so. The result of these new capital rules is that both banks and long-term savers have withdrawn from buying when others are selling, especially when markets become more volatile. In other words, financial markets have become less liquid.

The space previously occupied by bank traders has been colonised by algorithmic traders. “Algos” use computers to execute trades automatically and at high speed. But their defining feature is their slim levels of capital and their ability to change their trading strategies. In quiet times, algorithmic traders are contrarians: buying when the market is falling and selling when the market is rising. Their behavior adds to liquidity. When the market starts to trend, the algos switch into trading faster than others or more aggressively in the same direction of the trend. Selling more or before others drains liquidity and can create a flash crash. Algos generate fake liquidity: it is there when you do not need it and disappears when you do.

We need to incentivise real liquidity. We need taxes on very high frequency trading and to set capital for insurers and pension funds against their long-run risks rather than short-run volatility. But there is a bigger point to be made: the structure of the financial system matters. Bank regulators were right to demand that banks held more capital against the risks in their trading books than before, but they were wrong not to consider who should take their place in providing market liquidity. What this and the chase for the rogue traders reveals is that we are still trying to curb systemic risks by focusing on individual risks one at a time, and not thinking sufficiently about the system as a whole.