Challenge phases Main content User comments entryt's statistics and author info Related themes, inspirations and entries Share entry Challenge activity feed Footer links
Login

The Challenge

5502 followers

How might we improve the way citizens and governments interact? Read the challenge brief

entry

Inside the Vault: Interactive Citizen Access to U.S. Government Data on Financial Institutions

The US government has been collecting electronic documents from banks for years. We will explore how journalists and others can analyze this data. Transparency is the key to a vibrant financial system, and to preventing the next financial crisis.

Bank regulation is boring. Or, it used to be, until the financial system nearly collapsed in 2008, the European Union had a debt crisis, and banks in tiny Cyprus threatened the world economy. 

In other words, bank regulation is an issue we should all care about, because problems in banks affect the entire economy. Most will agree that information and transparency, combined with expert analysis, is the way to ensure a safer future for the financial system.

It turns out that much information is already available—if, by “available,” one means sitting in obscure files on government servers run by agencies most people have never heard of, and accessible (at best) through expensive third-party resellers. The US government is already doing a big part of the job: collecting data, requiring reports, using that data to make regulatory decisions, and making the raw data available to the rest of us.

The regulatory uses of that data are less transparent, however. The Treasury, the Fed, and other regulators no doubt use that data to inform the regulatory processes. But the regulators ask the questions within the bounds that Congress sets by law. We want to create independent assessments of the health and risk of the financial system.

It’s also difficult for the average citizen to make sense of the banks’ complexity in today’s financial system. Although the typical savings account is insured by the FDIC, a bank failure is still more than inconvenience and bailouts can be catastrophic. More to the point, systemic failures, as in 2008, can affect the entire economy, putting much more than a savings account at risk. And, of course, the American taxpayer ultimately pays the bills for bailing out financial companies.

In other words, the smart banking customer (and taxpayer) today needs to think about a bank not just as a safe place to put money, but more as an investor might evaluate the future prospects of the bank. We turn to the data as one of the best ways to make that evaluation.

In this project, our goal is to create public insight from information the government already has. We can take that raw data—quarterly filings by thousands of banks—and make it available for exploration on the Web. We propose to develop a continuously updated view (with some lag time, of course) that enables citizens to ask how safe their banks are, gives journalists the tools to dig into the data behind the stories they are writing, provides the raw material for independent academic research on the health of the financial system, and creates an open backstop to the Dodd-Frank safeguards. Building on that, we plan to correlate the bank filings with other public data to build a more capable system for analysis. Making the data more easily available has the potential to transform current research in these areas.

That’s only the beginning. Having the data available for analytics, we want to explore what we can really learn from it. Can we find the warning signs of future financial crises before they drive the economy off a cliff? As citizens and journalists, can we assess the efficacy of financial regulations? Can data scientists mine the available data to find the anomalies and piece the clues to hidden problems? Can we use the data to create a big picture of systemic risk? Can citizen access effectively supplement the work of the new Office of Financial Research?

We don’t know the answers to all of these questions. What we do know is that step 1 is bringing the raw data out into the light where we can look at it together, and begin to explore its utility.

This project brings together a novel collaboration of three research groups at Boston University: the New England Center for Investigative Reporting, the Center for Finance, Law & Policy, and the Hariri Institute for Computing and Computational Science & Engineering. Together, these groups bring together not only computer scientists and engineers to handle the technical aspect of data management and data analytics, but also the first customers for the data: journalists and policy researchers who are already interested in understanding the questions suggested here.

Beyond this initial group, an informal survey of journalists suggests that they are aware that the reports exist, but they do not use them—in large part, because while the data is “available,” it is not in a form for immediate use. We also believe that citizens want to know what the government is doing to prevent future financial crises, but they don’t have the information or expertise to make that assessment. By making the data more available, and connecting it to experts for analysis and interpretation, we can build that trust and enable citizens and experts to hold regulators, legislators, and bank management accountable for their oversight of the financial system.

As far as we know, there is no open data project focusing on available financial regulatory data. To be sure, some companies use it for their own analysis, but the results of that, and the ability to interrogate the data directly, are not available to the public. Our work aims to make the data and analytical tools openly accessible.
In this project, we propose to do the following:

  1. Build a web site and database for exploring the available data, along with a pipeline for importing the basic FFIEC data on a continuing basis.
  2. Create a simple scorecard of bank health for bank customers based on metrics from the reported data.
  3. Correlate the data with data sets available from other sources.
  4. Explore analytical models for assessing both the health of individual banks and risks to the banking system as a whole.
  5. Write descriptions to accompany the exploration tools to help citizens, journalists, and researchers understand what data they are seeing and where it comes from.
  6. Create online educational modules and integrate these modules in various open online courses, ranging from courses that demystify the data analytics for outreach and general education purposes, to courses that train journalists and academics looking for deeper understanding.
  7. Get feedback from the community on what questions need to be asked and answered, and respond to that as much as we can.
  8. Make the core software available to the community as open source.
  9. Create programming interfaces to the data so that others can write their own software to take advantage of the raw data.

Our goal is to make the web site open for use as soon as it is ready, so anyone interested can start using it right away. By incorporating community questions into the project, we hope that fast feedback lets us learn faster and deliver more insight.

The project team includes:

  • Maggie Mulvihill, Co-Director of the New England Center for Investigative Reporting at Boston University
  • Joseph Bergantino, Co-Director of the New England Center for Investigative Reporting at Boston University
  • Cornelius Hurley, Director of the Boston University Center for Finance, Law & Policy
  • Azer Bestavros, Professor of Computer Science and Director of the Rafik B. Hariri Institute for Computing and Computational Science & Engineering at Boston University
  • Win Treese, Associate Director of the Rafik B. Hariri Institute for Computing and Computational Science & Engineering
  • Boston University students and staff software developers 

For the first phases of data management, web site development, bank scorecard, and simple analytics, we are requesting $500,000. The funds would be used to support the development work, operations of the computer systems, students doing the initial analytical work, and development of the educational modules and courses.

We anticipate using the platform as an enabling part of further research proposals to support deeper analytics and academic research on financial regulation and systemic risk, as well as to sustain long-term availability of the data and web-based analytical tools.

Bank regulation may seem boring, but not to everyone. One of the biggest social lessons of the Internet is that it can bring together very deep interest and expertise that seems quirky in isolation, but powerful when brought together. Anticipating the next financial crisis may be the regulatory responsibility of the Financial Stability Oversight Council (the FSOC, created by Dodd-Frank), or it may fall to millions of users who are bank customers, investors, and counterparties.

We now know what failures of financial regulation can cost. This project is one investment in trying to avoid the next crisis.

What is your project? [1 sentence max]

We will Investigate current uses of bank filings with the US government and how this data may be used more effectively to understand bank health and systemic risk by building a set of web-based tools for collecting, analyzing, visualizing, and correlating the data with other information.

Where are you located?

Boston, MA USA

How did you hear about the contest?

  1. Email from Knight Foundation
Win's profile photo
entry submitted by: Win Treese
March 19, 2013, 04:55PM
63 views 0 comments 3 applause Applaud

Comments

Join the conversation and post a comment.

close

Login

Forgot my password?

New user? Sign up!