Trading costs for financial institutions are very high. Sometimes, a portfolio of well selected securities that would otherwise produce high returns, in fact ends up with subpar performance because the execution and implementation of their trading strategies is way too expensive. A great deal of research exists on the performance of hedge funds, mutual funds, pensions and the like, however, the performance of trading desks, a key financial intermediary, responsible for trillions of dollars in execution, has largely been overlooked by researchers. Because of this, the authors of “Performance of Institutional Trading Desks: An Analysis of Persistence in Trading Costs” set out to examine data on both Institutional trading desks and their brokers.
The authors examine a large data set, created with 48 million tickets, containing stock identifiers, which allow for gathering ...view middle of the document...
They then examine the relationship between trading costs and institutions abnormal holding period returns to determine whether institutions with high trading costs are those that are sacrificing trading cost for the opportunity take advantage of valuable private information. Next the authors control for the quality of institutional trading desks to isolate the contribution of brokers to overall trading costs thereby determining whether or not brokers can consistently provide low cost execution. Finally, they classify brokers into either execution-only or full-service categories and separately examine trading-desk performance to determine whether high cost, full-service brokers actually provide valuable research or favorable IPO allocation that translate to better portfolio performance.
There are some issues that remain unaddressed. First, the sample period ends in 2008, a year when many intuitions were forced to close doors and liquidate. Ancerno did not retroactively delete data for institutions that no longer existed in December of 2008. Trading costs would certainly be higher for institutions facing liquidation. Second, there is a selection bias in the institutions whose data are available to Ancerno. Only institutions who care enough about execution to hire a consultant, are included in the data. Perhaps this suggests the data largely represents institutions who trade relatively frequently such as hedge funds as supposed to long only funds. Or the data may only represent institutions who make very large trades and are therefore more concerned with costs. Furthermore, the data may be biased toward funds who seek to trade in a short time frame, such as event driven funds that cannot afford to wait for the cheapest possible execution strategy. There also may be a bias toward funds who trade in highly illiquid assets for whom trading costs are much higher. However, the data is only on equity trades, suggesting that liquidity, at least for the individual securities involved in the trade, is relatively high.