Pricing firms and bond indices often value corporate bonds inaccurately, negatively affecting trading books, dailymarketing to market, and risk management. This happens because bonds trade infrequently and portfolio values have to be priced using antiquated matrices and stale data.
To solve this problem, we developed the Repeat Sales Bond Index (RSBI). The repeat sales method can be the most efficient way of computing returns in a market with low frequency of trading. The method is successfully used in real estate markets. Case-Shiller Home Price Index is the most popular repeat sales benchmark.
All content on this website is currently in development and represents on-going research. All indices are still being tested for accuracy and are in need of significant improvement. Corporate bond pricing engine is currently outputting dummy valuations and is nothing more than a UI prototype.
This 2017 working paper by Matthew Spiegel and Adriana Robertson lays the theoretical foundations for applying repeat sales methodolgy to computing bond indices. It also suggests a number of tests for evaluating indices' relative accuracy.
This 2019 paper by Jens Dick-Nielsen and Thomas Kjaer Poulsen has being an invaluable guide for cleaning enhanced raw data from the TRACE bond databse. It includes SAS code filtering out canceled or inaccuratly entered trades.