While pursuing his doctorate in economics at Columbia University, Oliver Giesecke came across several challenges as he was researching and collecting data on the muni market.
Either it wasn’t available or it was incomplete or at times, misleading. In an effort to provide more and clearer data on certain attributes of the muni market, Giesecke, now a research fellow at the Hoover Institution at Stanford University, has , which aims to address several issues that many in the market often acknowledge have been around for a while.
The municipal dashboard consists of two components: market data and issuer fundamentals. The market data aspect uses real-time information on trading to come up with spreads for states and local government issuers, while an issuers fundamentals shows financial ratios, as derived from balance sheets and income statements, he said.
“We thought that having a dashboard that is interactive, and that people can use, is the best way to do it,” Giesecke said.
A complaint often heard in the muni market is that it is “extremely fragmented,” with more than 50,000 unique issuers, sometimes governments do not all issue bonds directly under one name, instead using intermediaries, something Giesecke quickly learned during his five-plus years of research.
“That makes it very difficult to link a bond in the market to who has actually issued it, and so that’s been sort of perennial challenge in the market,” he said.
The municipal dashboard aims to solve three challenges: bond linkage, pricing and spread aggregation.
Giesecke uses eight terabytes of data of the MSRB to link the bonds to the issuers. He came up with a machine learning algorithm that essentially goes through each of these documents and extracts the pertinent information.
Then he can take all the securities associated with an issue to get a credit spread.
Additionally, he found that because about 60% of all muni securities have embedded optionality, he wanted to provide insights into how that might affect pricing those securities.
He built pricing software for munis, which is free to use.
“You need a sort of a whole set of software to simulate different price paths and then determine what happens to the bond at each path. Once you have this information, you are able to get the price back,” he said.
For the municipal debt instruments pricing model, he as a “full-scale quantitative finance model with more than 300,000 lines of code.”
He said it can be done for 99% of all trades that occur in the market.
After combing through all of the trades, and to combat outlier trades that might skew the spread, he developed a statistical outlier algorithm.
Credit spreads can be sorted by state and cities, and week-over-week changes and time series, yields and trading volumes for those states and cities, Giesecke said.
Reception has been positive, he said.
“I’ve been in contact with a lot of municipal market professionals in the industry as well as think tanks and they thought that is a fantastic contribution,” he said. “There are some companies that have reached out that want to potentially have a data stream where they can use my data on a daily basis.”
There is the possibility to do this for every municipal trade in the market.
Currently, the dashboard is capturing around 25% of the market, but he said there’s still room to expand what entities the dashboard covers.