Looking for Market Manipulation Amid High Volatility
Hiding Amongst the Noise – Looking for Market Manipulation During Times of High Volatility
Market manipulation has long been a concern for many firms – but what happens when trading volumes increase during times of volatility?
MAR 20, 2020
By Michael O’Brien & Alan Jukes, Nasdaq
Market manipulation has long been a concern for many firms – but what happens when trading volumes increase during times of extreme volatility?
While many firms have invested early in preventing market manipulation, there are still firms who view enhanced trade surveillance processes as a ‘nice to have’ or even firms who may have implemented an automated surveillance system, but have taken a ‘set it and forget it’ approach.
During periods of low market volatility, it is easy to be lulled into a false sense of security with surveillance processes that simply check the box, but how do these processes perform when faced with the reality of strenuous market conditions?
There is no doubt that during times of high market volatility, data volumes are substantially larger than normal, which predictably drives up the total number of alerts being processed on a daily basis. The CFTC notably recognized this by temporarily relaxing regulatory obligations (see recent news from the CFTC).
However, in a backdrop of rapidly swinging market conditions, how feasible is detection of something more sinister than abnormal deviations?
Taking a historical approach to this idea, one could almost certainly expect to see one, if not multiple, cases of manipulation stemming from this period. Looking back, one well known example of blatant market manipulation correlated to a period of extreme market volatility (some would even argue causation) is the Sarao case during Flash Crash of 2010, when over a trillion dollars dissipated in just over 20 minutes.
The Sarao flash crash case
Consider the evidence the UK Financial Conduct Authority (FCA) and the US Commodity Futures Trading Commission (CFTC) used in the prosecution of Navinder Sarao and the flash crash layering case. Several attributes of Sarao’s trading were flagged as indicators of market abuse, many of which the executing brokers failed to detect and query.
Sarao’s layering algorithms generated a large number of canceled orders compared to entered orders. Moreover, his orders were modified much more frequently than other traders’ orders.
Additionally, there was a high ratio of cancelled orders to trade executions. Of the tens of thousands of orders that Sarao entered over the period, about 99.7% were cancelled without execution, compared to about 49% for other similar types of traders.
There was an unusual concentration of order entries within certain time windows. Typically the layering algorithm concentrated its activity within short bursts of time, it remained inactive for a while, and then become active again. Further, Sarao’s orders that were resting in the order book at any one time represented a substantial portion of the supply and demand for that particular instrument. The trading often was being carried out in a major US equities index futures contract, the S&P E-Mini. At times, the aggregated volume of Sarao’s orders represented 40% of the order depth for one side of the market.
Sarao’s orders were for various lot sizes, and the layering algorithm orders were much larger than average. Specifically, his orders were for 504 contracts on average, whereas the average for other traders was seven contracts. It is alleged that this sent signals that triggered other participants to come into the market.
If there is anything that can be learned from historical events, it is that anomalous behavior is not something to look lightly upon,
During times of strenuous market conditions, market manipulators can take advantage of the chaos and attempt to hide amongst the noise, assuming that their behavior will go unnoticed through all of the other trading activity.
Surveillance analysts need to be wary of a default assumption that alert spikes are simply a reflection of increased market activity, or due to the ‘normal’ response of an algo to increased volatility.
Firms must be prepared to challenge any trader that claims their behavior was due to abnormal market conditions, and the key to being able to challenge that is to look at their behavior relative to their peers.
When taking a closer look, there are certain behaviors that stand out from the crowd during these times, and that is still observable – even in times of wild volumes and high volatility.
Nasdaq spends 40 out of 52 weeks testing systems for the 120 marketplaces it supports.
Vehicles enable managers to tap the liquidity and tax advantages of ETFs, without giving away the strategy.
First-quarter ADV increased 45% from the first-quarter of last year.
There is appropriate protection for the development of Spanish secondary markets and financial systems.
Demand for RFQ services in cash equities has risen in current markets.