Can Social Media Help Predict Movement in Municipal Bond Market Prices?
April 19, 2022
By Christopher J. Mier, President & Founder, Rosebud Strategies LLC
The use of social media data is now commonplace for consumer products and eCommerce companies. Consumer products and internet-based companies routinely use social media analysis to better inform their business decision-making. With data drawn from the Internet from Blogs, News, Twitter posts, Instagram, Facebook and other channels, companies use social media data to understand how their brand is faring in the marketplace, what customer likes and dislikes are, and indications of the presence of brand loyalty. The effectiveness of marketing campaigns is often assessed through social media data.
With such widespread acceptance of the use of data culled from social media for consumer products, it’s no surprise that there has been growing interest in using similar data to better understand emerging developments in the municipal bond market.
One of the first financial sentiment indicators constructed from social media data specifically for use in the municipal bond market is the Rosebud Municipal Credit Diffusion Index (RMCDI).
Through the company I founded, Rosebud Strategies LLC, I created the Rosebud Municipal Credit Diffusion Index in January of 2022 to gauge the amount of negative sentiment surrounding state and local government finances in order to help inform municipal market participants about prevailing sentiment regarding the credit quality of this sector.
A rising Rosebud Index value means that the number of negative mentions regarding state and local government finances relative to positive mentions has grown. This, I’ve theorized is a clear – and generally unused – signal that there is an increase in the magnitude of broad concern regarding municipal credit quality. As a result, a rising Rosebud Index value should be correlated with rising credit spreads in the municipal bond market, which translates into higher yields and lower prices.
That at least was the proposition. But tests were necessary to see if it was a valid one. To that end, I tested to see if the overall measure of net negative sentiment found in mentions generated on Twitter were genuinely a valid approach to gaining new-found insights into market yields and prices.
Needing a good barometer of the overall municipal market to compare the RMCDI to, I chose the iShares National Muni Bond ETF (MUB), which is the largest municipal market ETF at $24 Billion, to see if there were statistical associations. The MUB is intended to track the investment grade municipal bond market, and is a reasonable proxy for the municipal bond market as a whole.
I tested for statistical relationships between the daily price movement of the iShares National Muni Bond ETF (MUB) and the RMCDI using the Granger causality test. This is a statistical means for determining whether a one time series is useful for forecasting another. The concept of Granger causality sidesteps the more complex issue of whether one variable can be truly said to cause movements in another. In this instance, however, the concept of causality didn’t seem to apply. At no point, did I speculate that Tweets somehow caused motion in the bond market. I simply wanted to see if they could help us to forecast that motion.
To my great satisfaction, when we applied the Granger Causality Test to the Rosebud Municipal Credit Diffusion Index, we found strong evidence that there were some positive correlations. The Rosebud Index, for example, was found to be highly correlated to activities in the municipal bond market over the period from March 16 2020 to March 18 2022, though it was seemingly less effective in the period prior to the pandemic.
Clearly, it will be necessary to continue to test the value of the Rosebud Index over time, to see if it continues to be a valid predictor of movements in the municipal bond market. But there’s little reason to believe that won’t hold true. Considering the vast amount of data that can be culled from social media, there’s every reason to anticipate that it will reflect consumer sentiments in this field.
As this blog item is being handed into the Government Finance Research Center, Elon Musk is attempting a hostile takeover of Twitter. To the extent that he succeeds, and alters the way that Twitter operates, there could be difficulties in comparing the way the Twitter feed correlates to bond markets. But notwithstanding that possibility, it is abundantly clearly that the sentiments about the future of government finances are a potentially useful tool to anticipate municipal bond market movements. And the social media are a powerful new way to judge the state of those sentiments.
The contents of this blog post reflect those of the author, and not necessarily those of the GFRC.