Can we agree that Market Impact is a bad thing?
Perhaps we can start with the premise that every buy side institution should strive to lower its Market Impact when executing its FX trades. No argument there, we hope. A recent article on market liquidity quoted one market expert who explained that the health (liquidity) in any marketplace can be measured by the size of the execution that causes the market’s bid or offer prices to move. This seems like a good corollary to the definition of Market Impact. Despite micro arguments to the contrary, the sheer size of the FX marketplace should allow most buy side institutions to execute trades without causing significant Market Impact – obvious exceptions occur when market volatility is impacted by news and other relevant macro events. Trading style, venue choice, and execution speed can all cause negative Market Impact. Those factors are all within the control of the FX buy side trader. When any of the factors create Market Impact, it costs the buy side money, plain and simple. Since the buy side are often directional traders, negative price movement caused by any of these factors is a compounding dilemma as the market is likely to continue to move against the trader until the volume is cleared and for a time thereafter.
Disclosed trading via the phone can cause Market Impact
Factors driving Market Impact
Here’s a simple fact – disclosed trading can cause Market Impact. Once the buy side discloses its intention to buy or sell a currency, the counterparty to that trade, with possession of that information, is free to adjust its prices – or “in-source liquidity” – to reflect the anticipated trade. If the buy side’s intended trade is large, the counterparty’s actions, based on its new knowledge or even educated guess, can cause more significant Market Impact. This type of disclosed trading intention occurs when the buy side uses RFQ’s or executes using “voice.”
Market Impact can also occur when the buy side faces market makers that employ last look. When a trade is rejected because of last look, the rejecting counterparty changes its pricing based on its decision to reject. The buy side, still needing to execute its trade(s), now faces a market that recognized its trade intention and has moved against it. Execution speed can also cause Market Impact. If the buy side institution approaches the market with a larger execution and attempts to clear that trade rapidly, it can disrupt what might be otherwise normal market conditions.
Deciding to use an aggressive versus a passive execution approach can also result in Market Impact. It benefits the buy side trader to know the pre-market condition of the marketplace. An over-bought or over-sold market will be further negatively impacted if the trader decides to employ one style or the other without such pre-trade market knowledge.
In the current algorithmic environment, buy side traders should also know that a poor distribution of child trades among counterparties can also lead to negative Market Impact if that distribution is concentrated among one or two market makers. That concentration will lead to price changes and information leakage. Today’s electronic market is a “connected” market. Major market makers are omnipresent across all venues. Executions, wherever they occur, impact price-making decisions. Because of that, the buy side must avoid any trading decision that allows market makers to easily read their intention and cause Market Impact.
When the buy side chooses to execute using algorithms, the measurement of Market Impact is more complex
Measuring Market Impact
Perhaps the best place to start is looking at execution style: risk transfer versus algorithm. The typical risk transfer trade occurs when the buy side trader uses either an RFQ or trades directly with a counterparty using “voice.” As we have argued, showing your name and requesting a quote creates Market Impact even before the trade occurs. In the electronic world you can measure that by looking at the price movement between sending the RFQ and executing the order. If the execution amount is your full amount, you may not care if the market moves after you are done with your execution. But if you have more to do, that’s another story.
When the buy side chooses to execute using algorithms, the measurement of Market Impact is more complex. We believe a good way to measure Market Impact is to compare the movement in rate between the first child execution and the average price for the parent order. Another alternative would be to compare the rate movement between the execution of the first child order and the execution of the last child order. But the later approach is a non-volume weighted approach – thus not reflective of real Market Impact. FX market makers have used a similar measurement to evaluate their performance. But for the buy
side, this approach does not provide the type of information that allows them to directly assess their execution choices versus their Market Impact. Venue choice and the presence of last look on those venues also complicate the measurement of Market Impact.
Executing on a last look venue can create Market Impact that is hard to quantify. To be meaningful, the buy side would require detailed reporting about “what rate did I miss” and “where did I finally execute.” Even that reporting would not paint a complete picture since the last look rejections would clearly influence all future executions as well. Since the goal is to measure Market Impact to improve trading decisions, buy side institutions executing trades on last look venues are effectively stymied when it comes to measuring Market Impact.
We think that current TCA products lack the functionality to measure Market Impact in a meaningful way. The best solutions in the future will require the use of far more data and will necessarily be more complex than any approach being touted today. Next generation TCA will feature analytics that can measure the market environment at the time of execution and calculate the impact of the execution on that environment.
