Nearly a year after the "flash crash" of May 6, 2010, the Securities and Exchange Commission (SEC) proposed a "limit up-limit down" mechanism that would limit the trading prices for listed equity securities to within a range near recent prices -- effectively limiting the realizable volatility of the price movements.1 The proposal called for price bands around the average price over the preceding five-minute period and would prevent execution of trades outside of these bands. The proposal was designed to circumvent trading errors and to damper market volatility.
ETF sponsors quickly backed the new SEC proposal since exchange-traded funds (ETFs) "have continued to be stung by flash crashes in single ETFs." An associate director of global ETF research at Morningstar said that the "proposal would provide a better experience for investors by avoiding broken trades and other inefficiencies."
A year later, the SEC approved the "limit up-limit down" mechanism with a few alterations to the original plan. The price bands were defined as "5%, 10%, 20%, or the lesser of $.15 or 75%, depending on the price of the stock." The first phase of the implementation essentially covered the securities in the circuit breaker program while the second phase would apply more widely.
It turns that out that one major problem with the program is the definition of the word "price". For securities that don't trade frequently, there may not even be a trade in the last five minutes. If there are no trades, you'd have to look at orders. As explained by Ugo Egbunike at IndexUniverse, whether an ETF's trading is halted under the new rules depends on the 'national best bid' (NBB) for that fund, not the price of executed trades.
The NBB was designed to give traders a sense for prevailing market prices; however, some have argued that high frequency traders can manipulate the NBB, reducing its usefulness for practical or regulatory purposes. In fact, there is some evidence that when 50 ETFs were halted on August 19, 2013, it was due to high frequency trading algorithms manipulating the NBB -- not as a result of trades being executed at large price deviations. While it is not clear why there were so many halts (or indeed how many is too many), there are some features of the program that could be susceptible to some forms of high frequency trading.
The New York Stock Exchange (NYSE) lifted the limit rules yesterday on over 500 ETFs because of their relatively infrequent trading (an average of 10,000 times per day or less). For these ETFs, the quotes on which the limit-up/limit-down system is based "can zip around based on available liquidity."
It will be interesting to see whether the SEC or the NYSE change the proposed rule to look at actual trades rather than NBB quotes (more akin to the previous system). While actual trades in relatively illiquid ETFs may be infrequent, it may be less susceptible to manipulation from high frequency traders.
_______________________________________
1This proposal expands upon the 18 month trialsingle-stock circuit breaker programintroduced by the SEC in June 2010 immediately following the flash crash. This program applied to S&P 500 stocks, Russell 1000 stocks and to a list of exchange-traded products that track stock indexes.