Current projects of the NLP::FIN::LAB
The following are some of the current projects our team is involved in. By and large, it is an eclectic
group, with somewhat of a focus in Finance (and Accounting and Business more generally), but also interests
in real estate, politics, innovation, and chess.
In terms of textual corpora we look at, it includes EDGAR, newspapers, earning announcements, analysts reports,
academic CVs, patents, accounting information, including extensive work on images and optical character recognition
techniques, as well as different types of scraping algorithms.
See the "Research output" page for more details on recent working papers
and published work by lab members.
- Working with images from the New York Times, we are scraping bond price
yields to study the term structure of interest rates precisely at the time the US Federal government started to rely
heavily on bond markets for its finances. We are also studying state and municipal bonds, which were a large
part of the bond market in the first half of the 20th century.
- We are studying the geographical location of industrial firms prior and during the Great Depression. Using the detailed information
provided in the Moody's manuals, we can track the geographical scope of industrial firms from today until the start of EDGAR (~1995).
Following prior work, we are using these manuals to also try to get at the network of managers and directors at US industrial and banking firms
in the first half of the 20th century.
- Our team is also studying academic output, in particular research papers and their impact in the field as a function of
different cross-sectional variables. While in a different playing field, this complements other research from the lab's team
on innovation at the corporate level, which also uses NLP techniques (LDA in particular).
- We are using high quality granular housing transaction and financing data at Zillow, merged with parcel data on renovations, to analyze entry/exit dynamics of housing liquidity
providers (e.g. cash buyers, flippers, investors, etc.) across times/regions and the implications for housing price/volume
dynamics. Our key interest is determining whether liquidity providers exacerbate or stabilize market turbulence.
We are also examining the role of investors in catalyzing neighborhood gentrification. We highlight tipping points and how
much neighborhood residents versus outside investors contribute to larger trends in neighborhood gentrification.
- Accounting statements were quite eclectic prior to the establishment of the SEC (and later FASB). Using the Moody's manuals,
our lab is trying to understand how reporting of both income statements and balance sheet information changed around
the introduction of the SEC, and all the new regulations that were passed in the 1930s.