Finding Just the Facts in Government Data

Former Microsoft CEO Steve Ballmer's latest venture offers an interesting twist on the aged argument that government should be run like a business. I'm sure you've heard that point of view multiple times in every election season of your adult life.

Most invocations of the "government as a business" pledge focus on how a candidate for office plans to slash government spending and make government many times more efficient. Of course, the candidate who makes that pledge soon becomes the politician who discovers that cutting costs often means cutting services that certain voters want, as well as the fact that government isn't a for-profit business. Said politician then complicates matters by sponsoring or enacting new programs -- promoted by campaign contributors -- that actually add costs.

Everyone wants to see government spending -- and taxes reduced -- but they don't want the spending cuts to impact them. With that in mind, Ballmer takes a new tack, focusing on using business concepts to gain a better understanding of government finances.

He announced USAFacts Institute, a partnership with the University of Pennsylvania's Wharton School, Lynchburg College, and the Stanford Institute for Economic Policy Research. Ballmer reportedly invested $10 million in the project.

USAFacts hopes to enhance our understanding of what the US government spends, where the money comes from, and where the money goes. The work is based on two well-known but very different types of documents. First, it draws on the language of the preamble of the US Constitution. Second, it is modeled on the US Securities and Exchange Commission's 10K form, which you probably recognize as the detailed annual report form for public companies.

On it's website, USAFacts says the venture had its roots in a discussion between Ballmer and his wife about where to invest his money. He set out to learn more about government spending:

"We soon discovered that dealing with something as big and complex as government – with its more than 90,000 jurisdictions and 23 million employees – required an organizing framework. What better place to look than the Constitution, and, more specifically, the preamble to the Constitution? It lays out four missions: 'Establish justice, ensure domestic tranquility; provide for the common defense; promote the general welfare; and secure the blessings of liberty to ourselves and our posterity.' While we don’t make judgments about policy, we all agree on the broad purposes of government as laid out in the preamble to the Constitution."

USAFacts already has lots of data points on its site, each associated with one of those tenets of the Constitution. For example, there is data about how much the average person might pay in taxes, how many arrests the police make, and how many visas are issued to foreign visitors. The logical next step is to assemble and organize enough data to help decision makers, and even voters, understand where we should be investing and where we should be trimming.

Ballmer, in an interview with Recode, said of today's mishmash of government data, "I think a lot of information is put out to make a point…People take a point of view, but then they pick the data that makes their point of view.”

For me, that statement about picking the data is that makes your own case seems to be evidence that government already is like business. It's exactly how business itself works. Corporate executives, department managers, even field staff like sales reps, increasingly have embraced analytics. Companies actually pick their preferred data points to convince customers to do business with them. Even those managers who don't want to hear about analytics have their own little data stashes in the form of experience. 

The problem is that too often business people only want the data that makes their point. That's the "good data." Others who disagree with them have "bad data." Then, just like government agencies, corporate types hoard their data because data is power. Those of you who gather and analyze data in the corporate realm refer to that as creating data silos. Have you noticed that every time we eliminate a data silo a couple new ones appear behind another department's walls? One result is that we never really have a good view of the corporate big picture because new data sources and new data definitions pop up even before last year's analytics project is completed.

That good data/bad data mindset isn't ready to change, whether in government or in business.

The Ballmer initiative is likely to do some good by highlighting a few possible action areas in government. But I don't hold hope that it will bring about major changes. I suspect USAFacts may end up pacing a step or two behind the government's ability to create data in new places. Even if USAFacts can identify the most current data that government generates, the odds are that it will fall victim to what nags the existing federal Open Data policy. Two departments with overlapping or interdependent functions will continue to use different data definitions, different data sources, even different time spans in their data reporting. Good luck consolidating that data.

Besides, in a political environment that accepts "alternative facts" or creates "fake news" we have to wonder which facts USAFacts will find to be, um, factual.