At our Financial Data Summit last month, we were honored to feature a plenary address by Alex Pollock, distinguished senior fellow with the R Street Institute, former president and CEO of the Federal Home Loan Bank of Chicago, a director of the CME Group, and author of Boom and Bust: Financial Cycles and Human Prosperity. He told our audience that “the time has really come” to standardize data across financial regulation. These are Mr. Pollock’s remarks as prepared for delivery.
One question underlying the very interesting data project and proposed legislation we are considering today is the relationship of data transparency to multiple perspectives on financial reality. In a minute, we will take up the question: Of all the possible views of a statue, which is the true view?
But first, let me say what a pleasure it is to participate in these discussions of financial transparency; the new Financial Transparency Act, a bill introduced in Congress last night; data standardization; and of course greater efficiency— we are all for making reporting and compliance less costly.
Let me add to this list the separation of data and analysis, or what we may call the multiplication of perspectives on the financial object. The potential separation of data and analysis may allow a richer and more varied analysis and deeper understanding, in addition to greater efficiency, in both government and business.
As the new white paper, “Standard Business Reporting,” by the Data Foundation and PricewaterhouseCoopers, says: “By eliminating documents and PDFs from their intake, and replacing document-based reporting with open data…agencies…gained the ability to deploy analytics without any translation.” Further, that standardized data “will allow individuals to focus on analytics and spend time understanding the data.”
In historical contrast to these ideas, it is easy for me to remember when we couldn’t do anything like that. When I was a young international banking officer working in Germany, one day 4,000 miles to the west, back in the Chicago headquarters, the head of the International Banking Department had lunch with the Chairman of the Board. Picture the Chairman’s elegant private dining room, with china, silver and obsequious service. In the course of the lunch, the Chairman asked, “For our large customers, can we see in one place all the credit exposure we have to them in different places around the world?” Said the Executive Vice President, International Banking, “Of course we can!”
The next day, all over the world, junior people like me were busy with yellow pads and calculators, wildly working to add up all the credit exposure grouped into corporate families, so those papers could be sent to somebody else to aggregate further until ultimately they all were added up for the Chairman.
That was definitely not data independent of documents. Imagine the high probability, or rather the certainty, of error in all those manually prepared pages.
A classic problem in the philosophical theory of knowledge turns out to be highly relevant to the issues of data transparency. It is the difference between the real object, the “thing in itself,” and any particular representation or perspective on it. In philosophical terms, the object is different from any particular perspective on it, but we can only perceive it, or think about it, or analyze it, from particular perspectives.
Likewise, a reporting document is a composite of the data—the thing—and some theory or perspective on the data which form the questions the report is designed to pursue and answer.
Let’s consider a famous type of report: GAAP financial statements.
Somewhere far underneath all the high level abstractions reflecting many accounting theories are the debits and credits, myriads of them doing a complex dance.
I think of my old, practical-minded instructor in Accounting 101. This essential subject I studied in night school when I was a trainee in the bank. I would ride the Chicago L train to my class, my feet freezing from the cold draft blowing under the doors. But this lesson got burned into my mind: “If you don’t know what to debit and what to credit,” he said, “then you don’t understand the transaction from an accounting point of view.” This has always seemed to me exactly right.
Later on, in this spirit, I used to enjoy saying to accountants advising me on some accounting theory: “Just tell me what you are going to debit and what you are going to credit.” This usually surprised them!
I wonder how many of us here could even begin to pass my old accounting instructor’s test when considering, say, the consolidated financial statements of JPMorgan. What would you debit and credit to produce those? Of course we don’t know.
For JPMorgan, and everybody else, the debits and credits are turned into financial statements by a very large set of elaborate theories and imposed perspectives. These are mandated by thousands of pages of Financial Accounting Standards pronounced by the Financial Accounting Standards Board. Many of these binding interpretations are highly debatable and subject to strongly held, inconsistent views among equally knowledgeable experts.
Any large regulatory report has the same character: it is a compound of data and theory.
An articulate recent letter to the editor of the Wall Street Journal argues that “The CPA profession has made the accounting rules so convoluted that GAAP financials no longer tell you whether the company actually made money.” This, the letter continues, is “why companies are increasingly reporting non-GAAP. Investors are demanding this information. …Why should public companies not supply shareholders with the same metrics that the management uses?” Why not, indeed?
In other words, why not have multiple interpretative perspectives on the same data, instead of only one? This is a fine example of the difference between one perspective—GAAP—and other possibly insightful perspectives on the same financial object. Why not have as many perspectives readily available as prove to be useful?
We are meeting today in Washington, DC, a city full of equestrian statues of winning Civil War generals. (The losing side is naturally not represented.) Think, for example, of the statues of General Grant or Sheridan or Sherman or Logan—all astride their steeds. Perhaps you can picture these heroic statues in imagination.
I like to ask people to consider this question: What is the true view of a statue? Is it the one from the front, the top, the side (which side?), or what? Every view is a true view, but each is partial. Even the view of such an equestrian statue directly from behind—featuring the horse’s derriere—is one true view among others. It is not the most attractive one, perhaps, but it may make you think of some people you know.
Likewise, what is the true view of a company, a bank, a government agency, a regulated activity, or a customer relationship? Every document is one view.
Pondering this brings back a memory of my professor of 19th century German philosophy. “The object,” he proposed, “is the sum of all possible perspectives on it.”
Similarly, we may say that a financial structure, or a policy problem, or an entity or an activity is the sum of many perspectives on it. The ideal of open data available for multiple reports, analyses and purposes is a practical application of this metaphysical idea.
The ideal is not new. In 1975, I went to London to work on a project to define all the elements—what were supposed to become the standardized data—for characterizing all the bank’s corporate customer relationships. The computing technology expert leading the project convincingly explained how these data elements could then be combined and reordered into all the reports and analyses we would desire.
Then, as now, it was a great idea—but then it never actually happened. It was before its time in technical demands. But now I suspect the time has really come. Fortunately, since then, we have had four decades of Moore’s Law operating, so that our information capacities are more than astronomically expanded.
-By freeing transparent, open data from being held captive in the dictated perspectives of thousands of reporting documents,
-By saving data from being lost in the muddle of mutually inconsistent documents,
-Can we provide transparent data, consistently defined, which will promote a wide variety of multiple perspectives to enrich our analysis and create new insights,
-Not to mention making the process a lot cheaper?
This would be a great outcome of the project under consideration in our discussions today.