This week I studied the paper "Algorithmic Trading and Information" by Hendershott et al. It addressed the question that Andrew asked me to consider very well. That is how can we study and exploit the relationship between the market and algorithmic trading.
The research was done based on the data collected from the top stocks on the Deutsche Market. It conjectured and proved that Algorithmic Trading improves liquidity, has no impact on volatility, can be detected through fee reductions, has a more significant impact on human trading than vice versa and also that it is more close to the efficient price than human trading more often.
The paper used vector auto regressions, impulse functions and covariance matrices in its methodology to understand the relationship between the stock market and algorithmic trading.