INTRODUCTION
Much has been written about the tradability and investability challenges of broad-based fixed income indices that contain thousands of bonds. Traditionally, fund managers seeking to replicate the returns of such indices typically choose a sample, holding a small fraction of the bonds. Tracking error to the benchmark may be managed by matching key risk characteristics, such as credit rating, sector, duration, convexity, and yield to-maturity of liquid and available bonds.
In this paper, we examine liquidity criteria that can potentially be used in the construction of replicable and liquid fixed income indices. We used the S&P 500® Investment Grade Corporate Bond Index, which contains more than 4,500 bonds and can therefore face the aforementioned replicability challenges, to guide the development of a framework for forming a narrower, investable basket. We then compare the liquidity profile of the resulting index, the S&P/MarketAxess Investment Grade Corporate Bond Index, to the benchmark. In assessing the relative liquidity of bonds, we examine the strength of commonly used indicators, such as size and age, against Trade Reporting and Compliance Engine (TRACE) corporate bond transaction data. Additionally, we examine the persistence of TRACE-based liquidity signals over time to develop a systematic approach in constructing a tradable subset while controlling for turnover.
METHODOLOGY
Liquidity may be defined as the ability to buy or sell an asset within a reasonable period of time with limited price disruption. Several methods exist to assess an asset’s liquidity. A simple way to compare the liquidity of two bonds is through the use of TRACE daily volume data. The data represents the daily aggregation of each reported trade throughout the day. The existence of reported volume data can be indicative of the frequency of trading. For example, if a bond has volume data for 20 of the last 22 trading days, then it trades relatively frequently—nearly every day.
