@article {Sommer21, author = {Philip Sommer and Stefano Pasquali}, title = {Liquidity{\textemdash}How to Capture a Multidimensional Beast }, volume = {11}, number = {2}, pages = {21--39}, year = {2016}, doi = {10.3905/jot.2016.11.2.021}, publisher = {Institutional Investor Journals Umbrella}, abstract = {Despite its importance, there currently exists no universally agreed upon and adopted measure or model that adequately captures cost and time to liquidation in bond (over-thecounter) markets. To fill this gap, we reviewed 40 years{\textquoteright} worth of research and summarize our findings in this article. We claim that the lack of concurrence on a definition can be attributed to the lack of consistent methodology. Connecting the dots within the vast body of literature, we find the key ingredients of such a novel measure: Taking market impact models as a natural starting point and adding the necessary math to quantify the inherent uncertainty of such a measure. We further suggest that machine learning methods are a natural candidate to overcome the main obstacles in this process, as they can help extract useful information from the extremely sparse data that form the main difference between equity and bond markets.TOPICS: Fixed income and structured finance, big data/machine learning}, issn = {1559-3967}, URL = {https://jot.pm-research.com/content/11/2/21}, eprint = {https://jot.pm-research.com/content/11/2/21.full.pdf}, journal = {The Journal of Trading (Retired)} }