Scientific Beta

Using transparent and replicable estimates of stocks' effective spread, we compute transaction cost levels and net performance of investable smart beta indices. We show that the extra performance relative to the broad cap-weighted index does not disappear once we account for their transaction costs independently from the measure used to estimate the effective spreads. We show that our results hold across strategies loading on several factors and several geographical areas with different levels of liquidity. We also analyse the main determinants of transaction costs. We show that the effective spreads decreased considerably after changes in market microstructure led to increased competition. Finally, we show that the benefit from diversification of transaction costs are higher across regions than across factors.

It is well known that it is important to account for transaction costs in the evaluation of smart beta investment strategies. Indeed, these strategies have raised concerns due to their higher turnover and higher exposure to illiquid stocks compared to cap-weighted market indices. Higher transaction costs are a likely "side-effect" of smart beta strategies and the reasonable question is whether these side effects outweigh the benefits. However, surprisingly little is known about the magnitude of these costs. In this article, we draw on Scientific Beta research to provide estimates of transaction costs for Scientific Beta indices.

The objective of this paper is dual. Firstly, we provide estimates of costs that arise when replicating investable index strategies using transparent methods that are replicable by investors. We then use these estimates to study the net outperformance of these investable factor strategies in different regional markets with different levels of liquidity. Secondly, we analyse the variation of transaction costs and the different dimensions that investors may use to diversify the risk of unexpected shocks to liquidity.