How Effective are the Order-to-Trade Ratio and Resting Time Regulations?

I. INTRODUCTION

The Flash Crash on the New York Stock Exchange on 6 May 2010 caused the biggest intraday point decline in the Dow in its history (near 1000 points).¹ What actually caused or exacerbated the Flash Crash? The first trigger may have been the spoof orders in S&P 500 futures contracts submitted and quickly cancelled by the high-frequency futures trader Navinder Singh Sarao.² Sarao entered more than 85 spoof orders to sell E-mini contracts at different times on 6 May 2010, the crash day, which represented more than 20 per cent of all visible E-mini sell orders. Once other traders in both the futures and equity markets saw the large sell orders and declining prices, they panic-sold numerous E-mini contracts or equity shares and drove the E-mini futures and the Dow into the historical rapid fall. Sarao was arrested on 21 April 2015 and charged with spoofing market manipulation and other counts of wrongdoing. He pleaded guilty on 9 November 2016, as Department of Justice prosecutors closed the case of spoofing.³

The Flash Crash reminded global securities regulators and investors of the potential threat that high-frequency trading (HFT) can bring to financial markets. In response, the securities regulators of the major developed economies have proposed or even enacted several rules to curb this trading strategy. The European Union inter alia adopted the Markets in Financial Instruments Directive II (MiFID II),⁴ which came into force on 3 January 2018. MiFID II has added new regulatory measures to safeguard financial markets from excessive algorithmic trading activity and HFT. Two proposals targeting these types of trading are of particular interest. One is the maximum order-to-trade ratio (OTR) below penalty. The other is the minimum resting time (RT) of submitted orders before a per order charge. On the national level, the German regulator BaFin enacted the HFT Act in 2013. It introduced a penalty for high OTRs. France levied a 0.01 per cent tax if high-frequency traders cancel their orders within a half-second.⁵ Many other nations, such as Australia, Canada, India, Italy, and the UK, also acted in this regard.⁶

The aim of this paper is to assess the two regulations targeting HFT enacted by several European regulating agencies, the OTR, and RT regulations. We select a common manipulative HFT scheme, spoofing, to anatomize the key to the manipulation. We find that the key to spoofing is to take advantage of the incomplete order display. Therefore, spoofing is a form of order-based manipulation (OBM) (see Section II). Equipped with the theoretical understanding of the manipulative HFT, one can clearly see at what point the OTR and RT regulations target the cycle from order submission to order execution/cancellation. Subsequently, one can understand what kind of difficulties the two regulations bring to spoofing manipulators, thus, the effectiveness and limitations of the two (see sections III and IV).

II. SPOOFING IS ORDER-BASED MANIPULATION

Spoofing is considered as market manipulation in Europe.⁷ The core of this trick is manipulation of the order display in the limit order book, which is publicly accessible. The main purpose of this tactic is to accelerate price changes by inducing other investors to trade in the direction desired by the manipulator.

Why does spoofing exist? More than half of the financial markets in the world today use a real-time limit order book as a display system to facilitate trade.⁸ This system displays open interest only. In other words, it displays only submitted orders that have not been executed or cancelled. However, there is no way to know if those previously displayed orders have been executed or cancelled. This display is incomplete and provides opportunities for manipulators to misguide numerous market participants, as a manipulator placing large spoof orders or fake orders can create the appearance of rich buy/sell liquidity. Thus, numerous orders in the direction desired by the spoofing trader are induced quickly. These induced orders substantially push prices up or down. Subsequently, the spoofing manipulator closes his position by trading against the induced investors with a sure profit in most scenarios.⁹

There are two basic scenarios regarding spoofing. This paper illustrates them with long examples. The first type of spoofing is also called momentum ignition.¹⁰ Under this type, the manipulator has a long shareholding position. He expects to close the position at a higher share price. He places large limit buy orders just below the best bid, so they can be displayed to the entire market. Numerous investors see the large buy orders submitted by the manipulator on display but cannot see if these orders are executed or cancelled. The investors develop the perception that large buy volumes are entering the market. Many of them rush to buy and push the share price higher. The manipulator cancels the spoofing orders completely before the price increase takes place. Thus, the manipulator has no new shares added to the early established position and incurs no new transaction costs. He waits briefly for the share price to be lifted above his expectation and, then, places sell orders. He closes his position by trading against the induced investors at a sure profit.¹¹ In this example, the manipulator uses spoofing to lift the share price in his accumulation-lift-distribution scheme.¹²

Under another scenario, the manipulator seeks to establish a new long position. He expects to buy shares at a price lower than the current price. To push the price lower, he places large limit sell orders near the current price. He makes sure that these sell orders are displayed to the entire market and, then, cancels them immediately. Then, numerous investors are pressured to sell when they see incoming large sell orders. Their trading pushes the share price lower, and it reaches the manipulator’s expectation. Thus, he establishes a long position at a lower price by trading against the selling investors.¹³

Spoofing involves order display manipulation, but no genuine trades. This creates ambiguity to other investors who do not know if the briefly displayed orders submitted by the manipulator are executed or not. Some of these investors are induced to trade. Their trading, frequently in herd behaviour, accelerates price changes in the direction of spoofing orders. Supplementary Table 1 illustrates the complete cycle of OBM that features spoofing.

In summary, spoofing is essentially order display manipulation. It comes into existence with a high success rate, because the extant order display by exchanges involves an incomplete disclosure of order history. Only submitted orders are displayed to the market and cancelled or executed orders are not. The ultimate purpose for some high-frequency traders to use spoofing in an OBM strategy is to ensure profits by creating and exercising informational monopoly power in trading.

Understanding the manipulative HFT from a theoretical angle helps to uncover the key to its tactics. Such an understanding also helps to assess the OTR and RT regulations. Supplementary Figure 1 shows which link the two regulations target in the cycle from order submission to order execution/cancellation. The OTR regulation monitors the relationship between order cancellation and order execution. The RT regulation targets order display. In the following sections, we will assess each regulation so that one can understand how effective and insufficient the two regulations are against spoofing.

III. ASSESSMENT OF THE OTR REGULATION

The maximum order-to-trade ratio limits the number of orders that can be cancelled using the total number of submitted orders as the reference. The effectiveness of this regulation lies in reducing the frequency of order cancellations, thus containing spoofing occurrences. Since the regulation does not distinguish the intent of the order cancellation, it will limit, to a certain degree, some order cancellation aimed at spoofing or other manipulation purposes. Unfortunately, it will stop some order cancellation out of non-manipulative intent.

The regulation involves a serious loophole, because it does not consider the difference in price impacts owing to the different sizes of orders. That is, regardless of whether the order is of one share or of 10,000 shares, the regulation considers both equally. Thus, one can submit a small order to be filled but a very large order to be cancelled. Although the ratio can remain within the cap, spoofing or other manipulation schemes can still be executed without violating the regulation. The following numerical example further illustrates this loophole.

Suppose the predefined ratio is 100:1. That is, the maximum ratio of submitted orders to executed orders is 100:1 before the penalty. Before the regulation is enforced, the manipulator submits numerous orders of large sizes and quickly cancels them all without a single order intended for execution. The ratio of submitted orders to executed orders is unlimited. There is no transaction cost during spoofing. Now assume that a spoofing manipulator submits 1000 orders with the equal size of 10,000 shares, intended for cancellation, during a trading day. As the regulation is implemented and the manipulator tries to remain legal, he also submits 11 orders with an equal size of one share during the same trading day. In fact, he submits each small order followed by 99 large orders until the last small order, which is followed by 10 large orders. He quickly cancels each large order but allows the small order to be executed. Thus, the ratio remains just below 100:1 throughout the trading day, and, thus, he does not violate the regulation. However, his spoofing scheme works out successfully too. The only difference the regulation makes is the small transaction cost to the manipulator owing to the executed 11 small orders compared with no transaction cost prior to the regulation. Supplementary Table 2 shows the before-and-after comparison for more general scenarios.

In summary, the effectiveness of the OTR regulation lies in the fact that it can deter spoofing manipulators and block excessive cancellations. Since it can be computerized and prevents spoofing manipulation, it is much more inexpensive and efficient than the traditional legal approach. Its insufficiency lies in the fact that it cannot eliminate spoofing by acting alone. This is mainly because the spoofing manipulator can get around the regulation easily by adding proportional small trades to the large orders that will be quickly cancelled.

IV. ASSESSMENT OF THE RT REGULATION

The minimum RT regulation targets the speed of order cancellations. Since the minimum time length is predefined, such as 0.5 seconds, and is universal to any instrument, the effectiveness of the regulation depends on how fast a limit order near the best bid/best offer is executed. In other words, if the difference of the order’s price and the mid-price of the best bid/best offer is given, the more actively traded instrument yields the shorter time of order display before execution. The loophole of the regulation is that it does not consider the liquidity of the instrument to be traded. A spoofing manipulator can switch to less actively traded instruments to seek a longer RT than the minimum required by the regulation. Another loophole is that the manipulator can choose a larger gap in the bid or ask from the best bid/best offer for the spoof orders to prevent them from being executed before the minimum RT. However, there is uncertainty created by the regulation for spoofing manipulation. In either of the aforementioned two scenarios, the spoofing manipulator faces the risk that his large orders on display get picked off by other investors. A more conservative manipulator would have to stop spoofing in the market where the RT regulation is enacted. Compared with the OTR regulation, the RT regulation is more effective in the deterrence of spoofing, since the former does not create any uncertainty for the spoofing manipulator. Supplementary Table 3 shows the before-and-after comparison for a long spoofing scheme.

In sum, the RT regulation poses uncertainty for the spoofing manipulator, since he cannot cancel large spoofing orders as quickly as he chooses to, but only after the predefined time for the display of such orders. Furthermore, once the spoof orders are mandated to be on display, the speed from submission to cancellation of orders is reduced substantially, since a high-frequency trader can complete an order life cycle within microseconds before the regulation. Even more than slowing down the trading, the regulation creates the risk that the spoof orders on display will be picked off by other investors. The most important aspect is that the imposition of minimum RT can substantially reduce the incidence of flash crashes.¹⁴ Since the regulation can be computerized and is preventive, it is more effective and less time-consuming or human-resource-consuming than the traditional legal enforcement.

The authors thank Xin Yan for inspiration and the session participants of 2017 Western Economic Association International Annual Meeting, 25–29 June for helpful discussions.

Footnotes