The GSSM generates expected shortfall numbers, which are estimates for the average change in stock price a trader may create by executing an order. Two major enhancements to the GSSM include:

• improvements to the US equities model
  
• calibration of that new model specifically for US exchange traded funds (ETFs).

The key enhancement to the US equities model is the removal of market capitalization buckets. Stocks within the GSSM are no longer classified into small, mid, and large-cap categories, which can be extremely fluid in times of increased volatility. Instead, the model has been enriched with a more flexible functional form that captures liquidity differences across the market capitalization spectrum and allows the model to better predict trading costs.

The addition of an ETF-specific version of the enhanced shortfall model allows for better transaction cost prediction when trading ETFs. While the old model included ETFs as part of the estimation universe for equities, ETF model coefficients are now estimated separately based on a large sample of executed orders for ETFs only.

"While we continuously monitor the predictive ability of our shortfall model, the significant volatility and dramatically changing US equity landscape of the past year underscores the value of these changes, which we researched over an extensive period of time," said Ingrid Tierens, Managing Director and head of equity execution strategies, GSET. "With this new model, clients benefit from an enhanced ability to predict shortfall estimates, not only for single-stock and portfolio trades, but also for ETFs, which have seen significant growth over the past few years."

The GSSM is based on Goldman Sachs' own executed orders data, rather than publicly available tick data, and is re-estimated frequently to reflect recent market conditions and uses factor inputs that are updated on a daily basis. Both customers of GSET, utilizing its electronic tools and services, and Goldman Sachs traders utilize the GSSM.