For our quantitative equity strategies we aim to minimize trading costs. We have developed our own trading cost model to estimate pre-trading costs. Trade evaluation results shows that we have realized cost savings of over USD 40 million for our clients over the 2009-2014 period.
One key part of our research agenda in recent quarters is a focus on lowering transaction costs. At first glance, this might look like a relatively minor issue. After all, the impact of a single transaction on net performance is minimal. However, when all the individually modest costs are bundled together, they can take a big bite out of net performance. This makes transaction costs a major consideration.
Achieving the right level of transactions is a fine balancing act. On the one hand, portfolio turnover should stay as low as possible to keep costs down. On the other, turnover has to be high enough to allow our models to work effectively. In fact, the turnover in our portfolios is already low, thanks to the use of our proprietary portfolio-construction algorithm and strict buy/sell rules. But we want to do better: every basis point counts. Our aim is to adapt our models to take into account stock-specific transaction costs in a more nuanced way. For one thing, transaction costs are not the same for all stocks. For instance, the bid/ask spread, which is included in transaction costs, is usually higher in emerging markets than in developed markets. It is also usually higher for small caps than for large caps.
It isn’t just a matter of looking at the visible costs that result from a transaction, which include commissions and trading-related taxes, in addition to the bid/ask spread. There are also implicit transaction costs such as market impact. As we were not satisfied with third-party pre-trade cost estimation models, we have built a trade cost model ourselves which is calibrated periodically based on all Robeco trades in recent years. Especially in emerging markets it is absolutely necessary to have trading cost data that are relevant for the type of trades that we do in practice for our quantitative equity strategies. Third-party models typically assume standard trades in terms of size, timing and location that substantially differ from Robeco quantitative equity trades. An add-on benefit of having an in-house model is that it can be easily integrated into the Robeco systems that we use for our quantitative equity strategies. Third-party models may require uploading potential trades into their tool and subsequently uploading the output back into the Robeco systems. Time to market can suffer from such a procedure.
The Trading Desk analyzes the executed orders on a trade by trade basis. The performance of executed equity orders is also systematically analyzed on a quarterly basis by Investment Technology Group (ITG), a consultant specialized in trade cost analysis. ITG measures the performance of traders and broker-dealers comparing these performances with the defined benchmark.
Over the period from 2009 to the end of 2014, over 62,000 trade decisions of the Quantitative Equity Strategies have been measured by ITG. The total value of the measured orders exceeded USD 56 billion. The outperformance of the trading benchmark resulted in a cost saving of over USD 40 million for our clients. Both in emerging and in developed markets the commissions we paid are low compared with the industry average. In addition, the Robeco Trading Desk has succeeded in adding value versus the trading benchmark in each calendar year by executing the trades with below average market impact.
There is a cost associated to trading. Robeco is committed to keep trading costs as low as possible. Keeping commissions low, recognizing both explicit and implicit costs and implementing smart trading tactics has resulted in a clear added value for our clients in recent years. For more details, please read the white paper Operational excellence in quant equity trading below.