Do dark pools amplify volatility in times of stress?

ABSTRACT

In this article, we explore the relationship between trading on dark pools and equity volatility in the context of the recent concerns about increase in frequency of market shocks combined with changes in market microstructure. In order to understand the potential role of dark pools in times of stress and implications for financial stability, it is essential to investigate the relationship between investor trading preferences and market volatility. For our analysis, we use data on daily trading volumes of FTSE100 stocks on dark and lit order books. We find evidence that dark pool trading has explanatory power in predicting volatility, implying that dark pools may affect the dynamics of price formation through liquidity. Our findings suggest that increased use of dark pools does not increase volatility, but may in fact lower it. Thus, dark pools may not be significantly detrimental to market stability in times of stress. This highlights the need for further analysis of the effects that shifting financial market structure might have on financial stability.

KEYWORDS:

• Dark pools
• liquidity
• volatility
• financial stability

JEL CLASSIFICATION:

• G10
• G14
• G18

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ 28.6% of all volume traded on the 15 DOBs consisted of FTSE100 stocks, only 11.1% of DAX stocks and 12.9% of CAC40 stocks

² Volatility = (P_max – P_min)/((P_close + P_open)/2); P_max and P_min are the maximum and minimum index prices, and P_close and P_open are the opening and closing prices of the index, respectively.

³ A Dickey–Fuller test reveals that we cannot eliminate the possibility that the share of trading conducted on dark pools has a unit root.

⁴ The interaction term is x × i, where x is volatility or dark pool usage, and i is an indicator variable identifying days of high volatility (defined as higher than the 95th percentile of volatility in the last quarter) or high dark pool usage (a growth of over 1.95% in the share of dark pools, the 95th percentile of the days in the sample).