Bond Liquidity and Dealer Inventories: Insights from US and European Regulatory Data
Feb. 11, 2020
Plamen Ivanov, Alexei Orlov, and Michael Schihl
Access this paper via SSRN: https://ssrn.com/abstract=3529801
Most corporate bond research on liquidity and dealer inventories is based on the USD-denominated bonds transactions in the US reported to TRACE. Some of these bonds, however, are also traded in Europe, and those trades are not subject to the TRACE reporting requirements. Leveraging our access to both TRACE and ZEN, the UK’s trade reporting system which is not publicly available, we find an overlap of about 30,000 bonds that are traded both in the US and in Europe. This paper examines how using the CUSIP-level information from TRACE and ZEN affects the computation of bond liquidity metrics, dealer inventories, and the relationship between the two. We find that in the combined dataset, the weekly volume traded and number of trades are significantly higher than in TRACE: e.g., the average unconditional number of trades in investment-grade (high-yield) bonds is 17% (20%) higher and the average unconditional volume traded is 15% (17%) higher when we incorporate the information from ZEN. We find a strong positive relationship between inventories and liquidity, as proxied by the trading activity metrics (i.e., number of trades, zero trading days, or par value traded) in TRACE data, and this result carries over to the combined dataset. When measuring bond liquidity with the Amihud ratio, we find strong relationships in both TRACE and ZEN but of opposite signs: greater (lagged) inventories result in higher liquidity in the US but lower liquidity in Europe. The two effects offset each other and significance disappears in the combined dataset. We conclude that (i) neither of the individual datasets paints a complete picture of the effects of dealer inventories on bond market liquidity, (ii) the measures based on the combined dataset appear more precise in describing the market characteristics, and (iii) data sharing across transaction reporting databases would allow a variety of stakeholders to gain a more accurate understanding of the liquidity and dealer inventories in global bond markets.