Are Circuit Breakers Doing Their Job?

It has been so long since the stock market fell precipitously that most had forgotten that circuit breakers and price limits even existed, and fewer still remember why they were instituted in the first place. The Crash of 1987, though, was by far the largest, scariest one-day move in the stock market in modern history (even larger than the stock market crash of 1929). It was this event that led to the implementation of circuit breakers and the reignition of the discussion today about whether they serve their purpose.

Historical Context

A tad bit of history to set the stage: the Brady Commission (named after then-Treasury Secretary Nicholas F. Brady), which was formed to evaluate the 1987 Crash, decided that there were a number of causes of the crisis. One was the use of a strategy called portfolio insurance which required users to continuously adjust their S&P futures positions to maintain their protection. This hedging strategy had increased in 1987 in what was viewed as an overly exuberant equity market. Seemingly benign during the uptrend, the strategy requires increasing the number of short (selling) positions in the futures contract as equity prices fall. At the time, this dynamic hedging technique was viewed by the Brady Commission as a contributing factor that helped accelerate the down move on October 19, 1987—to the tune of nearly 23 percent that day. 

Even though subsequent studies have questioned the influence of dynamic hedging during the 1987 Crash, it brought to the fore an important element—markets are connected and any attempt to slow or halt one will inevitably lead participants to move to another. So, when the Commission recommended  instituting “circuit breakers” on the NYSE when the S&P 500 index declined by a certain amount, they recognized both the stock market and the associated futures market would need to halt trading simultaneously for a given period of time.

Of course, this coordination was not so easy to implement, since “cash” stock markets did not have any limits and futures exchanges had long used price limits to constrain the potential amount of daily margin payments, which require traders with losing positions to pay up based on daily price changes. The futures price limits were not time-bound halts but thresholds beyond which prices could not move. The circuit breaker idea was grounded in the notion that a time-out could allow participants to clear their heads and that a panic could be quelled. It was not expected that such a “time out” would necessarily alter the direction of prices, only that it might prevent overshooting. After the predetermined time, trading could resume with prices adjusting immediately to a new (possibly lower) level. In the end, the NYSE and CME, where the futures were traded, agreed on the circuit breaker concept for this market. Over the years, several adjustments were made to the amount of the declines and time frames for closure as the exchanges gained experience. 

Fast forward to 2020: the thresholds on the cash equity index market (NYSE) for market wide circuit breakers since 2013 have been a decline of 7 percent for “level 1” with a 15 minute halt, 13 percent for “level 2,” again with a 15 minute halt, and 20 percent for “level 3” with a halt until the end of the trading day.  The equivalent set of price limits exist on the Chicago Mercantile Exchange for the equity futures.  One twist is there is a 5 percent up or down set of limits for the overnight futures trading session.

What does previous research tell us?

With four circuit breakers hit in a little over the week of March 9-18, and the subsequent decline in prices and increase in volatility in the U.S. stock market, many are questioning their usefulness.  Do those trying to sell accelerate their sales orders to “beat” the closure when the price is headed down?  Or knowing that they will be stuck for some period of time that they avoid putting into the market orders that forces the limit to get hit. This notion of the limit as a “absorbing barrier” or a “magnet effect” rather than a “repelling barrier” has been tested econometrically.  Using mostly futures price data (Ma, Rao and Sears (1989), Kuserk et al (1989), Arak and Cook (1997)) the results are mixed, though more recent simulations using hypothetical prices calibrated to more recent periods (Lera, Sornette, Ulmann (2019)) suggest the “magnet” effect would predominate.

Since circuit breakers are so seldom hit, academicians have built theoretical models to examine the usefulness of circuit breakers, again with mixed results. Though it is fair to say that when compared to unimpeded trading most results suggest any stoppage (including circuit breakers or price limits) is detrimental as it does not permit prices to reflect all available information. There are cases, however, where limiting prices may be beneficial to society (Kodres and O’Brien (1994)). In general, however, stoppages make markets less “efficient”—often delaying price movement and suppressing volatility that comes later. Models that allow for price dynamics close to the threshold find that the magnet effect is more likely as prices approach the level of the halt or limit (Subrahmanyam (1994); Chen, Petukhov, Wang (2019)). Alternatively, when models incorporate more realistic attributes of trading and their participants (for instance, participants have differential access to information) as the benchmark, the results are more favorable to the use of circuit breakers (Greenwald and Stein (1991)).

One simple way of detecting whether circuit breakers are making a panic worse (the magnet effect) is to examine the price behavior in the minutes and seconds close to the circuit breaker as well as after trading resumes. In the latest set of circuit breaker “hits” it appears the price falls are quite large and accelerate toward the threshold. However, the NYSE reports that “On three of the four days, the market stabilized near the 7 percent down trigger for most of the rest of the day. On March 16, the market traded until noon in the neighborhood of the 7 percent-down trigger, but then later in the day traded down almost twice as much.” Although a counterfactual analysis is difficult at the moment, this evidence suggests that the circuit breakers slowed, but did not completely inhibit downward pressure and the magnet effect is in play. It is notable that the 13 percent trigger has not been hit despite a peak to (current) trough of around 33 percent.  

Another way to see if the magnet notion is more prevalent is to examine the types of trading strategies that could accelerate prices toward the barrier. The 1987 portfolio insurance strategy is just a standard options strategy and over the years similar strategies featuring “nonlinear” acceleration near limits have been noted. More recently some have questioned the use of automated computer-driven trading in this context (for instance, these new trading mechanisms received a lot of attention in the wake of the Flash Crash of 2015). As well, there are questions about whether Exchange-Traded Funds (ETFs) are also contributing to faster price moves toward the closure. Of course, unless one is continuously keeping track of these kinds of strategies and know how much of the volume they represent it is difficult to say anything a priori about whether any of these strategies would be (or are) hindered by circuit breakers. Still, it is reasonable to think that with the rapidity with which strategies can be executed in today’s markets, these accelerating features are more prevalent today than in earlier periods.

What purpose are circuit breakers serving?  

The latest round of circuit breakers provides an opportunity to reflect and potentially refine them. In doing so, new trading strategies and use of new products (e.g., ETFs) will need to be taken into account. Even if the original goals of circuit breakers—preventing overshooting price falls or quelling volatility—seem elusive, there are still at least two reasons to consider not throwing out the tool of circuit breakers entirely. One, even if it is unlikely that traders will change their minds about where prices are going, they might. By forcing a short closure, it allows both sellers and buyers to reassess their orders, and perhaps pulls in potential buyers. Two, it allows the markets time to process trades and the infrastructure to catch up. For futures markets, the break allows the clearing house to assess the credit worthiness of its clearing members and allows for interim calls for additional variation margin. Allowing the infrastructure time to work is actually very important to ensure participants are treated fairly—if there is even a hint of difficulty in execution then the original panic is combined with mistrust. This combination ultimately will be an even larger problem than the falling prices.

Bibliography:

Arak, Marcelle and Richard E. Cook, 1997, “Do Daily Price Limits Act as Magnets? The Case of Treasury Bond Futures,” Journal of Financial Services Research, Vol 12, No 1, 5-20.  

Chen, Hui, Anton Petukhov, and Jiang Wang, 2019, “The Dark Side of Circuit Breakers,” Working Paper, MIT Sloan School of Management.

Lera, Sandro Claudio, Didier Sornette, and Florian Ulmann, 2019, Price Dynamics with Circuit Breakers, Working Paper, available at https://ssrn.com/abstract=3143247

Greenwald and Stein, 1991, “Transactional Risk, Market Crashes, and the Role of Circuit Breakers,” The Journal of Business, Vol. 64, No. 4, 443-462.

Kodres and O’Brien, 1994, “The Existence of Pareto-Superior Price Limits,” American Economic Review, Vol. 84, No. 4, 919-932.

Kuserk, Greg.; Moriarty, Eugene.; Kuhn, Betsy.; and J. Douglas Gordon, 1989, “An Analysis of the Effect of Price Limits on Price Movements in Selected Commodity Futures Markets,” Commodities Futures Trading Commission Research Report no. 89-1. Washington, D.C.: Commodities Futures Trading Commission, July.

Ma, Christopher K., Ramesh Rao and R. Stephan Sears, 1989, “Limit Moves and Price Resolution: The Case of the Treasury Bond Futures Market,” Journal of Futures Markets, Vol 9, No. 4, 321-335.   

New York Stock Exchange, 2020, “Circuit Breakers are Doing Their Job But Don’t Close the Markets,” contribution by Nicholas Brady and Robert Glauber, March 2020, https://www.nyse.com/article/circuit-breakers-are-doing-their-job-but-dont-close-the-markets

Subrahmanyam, 1994, “Circuit Breakers and Market Volatility: A Theoretical Perspective,” The Journal of Finance, Vol. XLIX, No 1, 237-254.