I recently had the chance to visit a company that is engaged in providing Accounting Fraud Risk Metrics for public US companies. Even though many of the data analysis techniques are the same, their overall outlook is vastly different from what I’m used to, due to their exclusive focus on company balance sheets and publicly released accounting data. In the world of quant research, especially so for HFT, there is a huge disconnect between the fundamentals and the information used to devise actionable strategies.
While I was working in the quant investment sector, it was almost pointless looking at the company balance sheets. For our purposes, it didn’t matter whether companies were healthy or unhealthy – only their trading characteristics mattered. And since were were focused on the market as a whole rather than any single corporation, balance sheet analysis really never came up at all. Even if it had, I would take what was written in the public filings with a grain of salt, after hearing about the shenanigans of Lehman Brothers, Bear Stearns and all those other wall street jokers.
My boss often told me about an old strategy that focused on executive options grants to generate signals. What he found was that there was often foul play involved, where executives would backdate the options to a low point in the stock price, making more for themselves at the expense of the public. I assumed that was just one of the many accounting tricks available, so not being an accountant myself and not knowing how to spot fraud on my own, I haven’t bothered to look into fundamental data ever since.
The good news for me, and anyone else harboring a distrust for the accounting practices of large corporations, is that there are companies like Audit Integrity out there to let people know what’s what. They are well known in their niche for employing several automated systems to generate and distribute various accounting metrics to their clients. Using SAS and operating on a large SQL database of publicly available accounting reports, they have created a highly accurate fraud detection model that takes into account historical cases of fraud to weight specific metrics for their ability to predict whether a company is currently committing accounting fraud. Not only are these metrics calculated for each of 5000+ public companies, each metric is compared against that of all related companies and displayed in a distribution chart within the AI web interface. After combining all metrics to generate an aggregate score, companies are ranked by percentile as far as how likely they are to be committing fraud. Overall I was very impressed by their system and hope I get another chance to explore their data very soon.
Speaking of data, when I visited their office they asked me to think of ways to reduce their data load and streamline their workflow. Another issue they asked me about was how to break into the consumer market with their service offering. Without more details about the inner workings of their database system I can’t make any specific recommendations, but I did think a bit about how they might expand their business.
For a long time I’ve been thinking about possible projects that might make use of Microsoft Live Labs Pivot, and Audit Integrity is definitely a service that would benefit from using such a visualization technology. Currently their service allows companies to be ranked by percentile, specific companies to be viewed, and a few other specific searches. But it seems to me like it would take a highly sophisticated person or very specialized job requirement to make full use of the data they provide.
My suggestion to Audit Integrity is to take a look at implementing a Pivot style visual interface to their data. This could be offered in addition to their existing services, and also be used as a great marketing tool. I think that a lot of consumers nowadays (in this case retail investors) may be like me in that they are in love with interactive infographics and innovative presentation of complex data. Some of my favorite blogs are FlowingData, VisualComplexity, which both do a great job of showcasing this type of work.
So as long as the funding is there, and all their current operational requirements are going smoothly, I think a Pivot interface would be a great addition to Audit Integrity’s services. Aside from being great publicity, I think it is a better way of accessing the data that will allow their clients to actually get more value from the metrics that they generate. Pivot application data is all kept on a central server and only downloaded when a user drills down in more detail, so it may save bandwidth as well. And once all their data was in a format easily accessible by API, they could also sell access to hedge funds or institutional investors for use in automated strategies.
Link to official MS Press Release: New Microsoft Live Labs Pivot Technology Brings Information to Life