Skip to content

High‐Frequency Trading around Large Institutional Orders

Bibliographic record. Follow the original-source link for the publication.

Field Value
Primary domain Institutional Execution
Other domains —
Methods —
Facets —
Authors VINCENT VAN KERVEL, ALBERT J. MENKVELD
Published 2019-03-21
Source The Journal of Finance
Identifiers doi:10.1111/jofi.12759
URL Open original source

Editorial synthesis

Why it matters

Using granular institutional executions, the study shows how HFT flow can shift from early liquidity provision against a large order to later same-direction trading, and links that transition to order information and implementation shortfall, directly informing the trade-off among execution duration, information leakage, and order splitting. (full_text:S45, full_text:S49, full_text:S54, full_text:S69, full_text:S76, full_text:S83)

Main author claims

  • The authors report that HFTs generally trade against institutional orders at the start but turn to same-direction trading for orders lasting more than about seven hours; a placebo sample with similar market conditions but no institutional activity shows the against-wind pattern but not the with-wind pattern. (full_text:S54, full_text:S60, full_text:S63)
  • The authors report that stronger HFT same-direction flow is associated cross-sectionally with more informed institutional orders, with a one-standard-deviation increase in with-wind flow corresponding to a 9.28-basis-point increase in permanent price impact. (full_text:S383, full_text:S384)
  • The authors report that implementation shortfall is about 4 basis points below average with stronger against-wind HFT flow and about 9 basis points above average with stronger with-wind flow, while cautioning that endogeneity and reverse causality likely overstate the latter relationship. (full_text:S526, full_text:S527, full_text:S528)

Data, method, or discussion scope

The evidence combines 801,341 child trades from four large institutions with public NASDAQ OMX trades that identify counterparties, covering 30 Swedish index stocks from January 2011 through March 2013. The authors construct 5,136 metaorders and compare them with placebo periods having similar market conditions but no institutional activity. These results have not been independently replicated here. (full_text:S46, full_text:S47, full_text:S48, full_text:S49, full_text:S52, full_text:S60, full_text:S63, full_text:S154)

Main limitations

The authors explicitly make no causal claim; the main results are unconditional or regression associations. The data omit HFT trades on other venues and observe institutional executions rather than complete orders. The sample contains two mutual funds, one sovereign wealth fund, and one pension fund, but no stressed hedge funds. The authors also identify endogeneity and reverse causality in the execution-cost relation, so the associations should not be treated as a deployable execution rule. (full_text:S103, full_text:S104, full_text:S165, full_text:S395, full_text:S527, full_text:S528, full_text:S579, full_text:S585, full_text:S593)

Relationships

  • None recorded.