Financial Crisis: Why didn't they use analytics?

For the past while, I have been reviewing the aftermath of the 2007-2008 Financial Crisis. I came across an interesting piece that highlights the importance of using analytics and "dashboarding" to monitor risk within a company. To be specific, I came across this when going through Nomi Prins's book, It Takes a Pillage: An Epic Tale of Power, Deceit, and Untold Trillions. Nomi Prins was in charge of analytics at Goldman Sachs and other banks. The embedded video gives more information about her and the book she wrote:



While listening to her book, I came across a transcript from the hearings in the aftermath of the crisis. As can be seen in the following video, Representative Paul E. Kanjorski is questioning the now-former CEO of Country Wide financial, Angelo Mozilo about the sub-prime crisis.



The part to focus on is where he grills the CEO about why they didn't aggregate statistics to monitor the mounting losses from the sub-prime loans (click here for where the transcript was extracted from. Please note the italics and bold is mine):
"Mr. Kanjorski: How long did it take you to come up with the understanding that there was this type of an 18 percent failure rate before you sent the word down the line, "Check all of these loans or future loans for these characteristics so we don't have this horrendous failure?"
Mr. Mozilo. Yes, immediately--within the first--if we don't get payment the first month, we're contacting the borrower. And
that's part of what we do. And we are adjusting our----
Mr. Kanjorski. I understand you do to the mortgage holder. But don't you put all those together in statistics and say, "These packages we are selling now are failing at such a horrific rate that they'll never last and there will be total decimation of our business and of these mortgages?" "

In other word, the Congressman is wondering how the CEO could not know that his business was failings because it is only common sense to monitor the key metrics that measure the key risk indicators (KRIs) associated with his principal business activities.

I would be the first to argue that there was much bigger issues with the financial crisis, such as the 16 trillion dollar-bank-bailout, the failure to properly rate the bonds backed by the sub-prime mortgages, quantitative easing, and so on.  That being said, organizations and companies need to be aware of the importance of measuring the KRIs associated with their business. Regulators, and others charged with oversight, will eventually question the insufficiency of such monitoring controls. Furthermore, as these regulators are more tech savvy - such as the judge in the Oracle vs Google trial - the more sophisticated dashboards they will expect.

The Power of Visualized Analytics

In my new role at Deloitte, I have recently come across tools, such as Tableau or Qlikview, that allow users to "visualize data". To be honest I didn't think they would add much value compared to "rule-based analytic tools", such as IDEA and ACL. However, after using these tools I realized the real power of being able to visualize data in contrast to producing an exception. It brings the dashboard concept within the executive management suite to the analyst or other business professional. But as they say "seeing is believing".

So let's try an experiment.

I recently came across an amazing visualization that really illustrates that power of visualization that visualizes economic data, specifically the distribution of wealth.

But don't click on it yet!

To get the most of the experiment first read this report (which the visualization is based on) to see how the stats hit you in terms of impact.

So here is an excerpt from the Oxfam report which the visualization is based on. (The numbers at the end of the sentence are footnotes; see the original report for the sources)

And

So now let's see how this data (plus other sources) hits you when it is visualized:


Is there really any contest?

What I've realized is that the visualization really enables the business user to bring together multiple dimensions into a single sheet of paper and enables you to tell the story about the underlying data. Having said that, I do believe that there is a complementary relationship between visualized analytics and rule-based analytics. For example, if you want to quantify the difference between budgets-and-actuals, produce a list of exceptions, etc, then rule based analytics are better for such a purpose. Furthermore, visualizations can help explain the results of rule-based analytic procedures.