Lost in the firestorm over AIG bonuses yesterday was a quiet announcement by Attorney General Eric Holder that laid another solid plank in the government's structure of transparency and accountability. Michael J. Sniffen's AP story, "AG overturns strict Bush-era information disclosure rules," has a cutline summarizing the change underway in Washington: "Agencies urged to release records unless law forbids," i.e. there's an a priori bent toward openness. The AG was way ahead of schedule with this, acting earlier even than the May 21st deadline for policy directives implementing President Obama's Open Government initiative, established by the very first acts of his administration (see "Obama's Midnight Memos).
With the presumption of release of government information, transparency and trust take a big step forward. Which brings us to the second part of "Mapping the Recovery" (see part 1) in our NetAge Reports series. Here, we address the specifics of the data that will be used to provide transparency and accountability for the chunks (block grants) of spending, particularly of the $787 billion passed by Congress in February.
It is especially important that tracking the expenditures be done correctly and that it truly makes responsibility and accountability clear. Renewed public trust rides so heavily on how this money is spent. While AIG's shenanigans show how difficult it can be to engender trust at the recipient end of spending, particularly in the private sector, the government has an opportunity to increase trust in itself, in how it is handling this money, and making sure responsibilities are clear. Done properly, this builds social capital, quite literally, that will make other parts of the multiple critical agendas possible.
Geek (or wonk) warning: this part of the story, below the break, is more than a little geeky (and plenty wonky--a bit of each). We discuss the data that is required for reporting the recovery funds, then what's missing. Specifically, what's missing is the organizational chart of responsibilities--which is not so hard to fix.
We've created a prototype map of the network of relationships called for in OMB's recovery reporting requirements. We illustrate our example with a dummy grant at the end of a real recovery program, which we show in four views: the government agency responsible for the grant, the grant recipient (whom we're inventing in this example) who will be held accountable, the physical location of the grantee, and and the overall purpose of the grant program. You can see the recovery network model live in your browser (wait a moment for the Java to load), and trace the grant along four axis as described below.
-- Jeff Stamps
What’s in the Recovery Picture?
Accountability starts with transparency of details. What, exactly, will we know about each block of money? How will we find it, track it, and understand it?
In network lingo, the recovery data model laid out in the OMB memo specifies six nodes, including the grant node. Each node has a unique ID situated in a defined classification system, and it has a line of text that describes it, a memo field, if you will, providing a name or title. The six nodes are:
-
Grant block, a chunk of money with its record ID;
- Responsible agency, with a Treasury Account Symbol (TAS) code;
- Program vehicle, identified by a code in the Catalog of Federal Domestic Assistance (CFDA) which provides a wealth of purpose-related categories to sort grants by;
- Agency account, the extension of Treasury’s TAS code;
- Recipient of the funds, identified by a Dun & Bradstreet number (DUNS), which, like the program’s CFDA code, provides entry to an array of associated details and purpose-related classification systems (e.g., sector, industry); and
- Place, the physical location of the recipient, completely defined in the data model by an address with a zip code and a city, county, and state Federal Information Processing Standard (FIPS) code, which enable easy association with political districts benefiting from the funds.
The grant block record laid out in the initial guidance (see Figure 2) means, for example, we can follow a grant to a zip code, which puts this chunk of money in the context of a nested set of legal and political jurisdictions capped by states. Simultaneously, the same record can be represented in a variety of other classification systems, such as one organized by purpose (CFDA, the Catalog of Federal Domestic Assistance) or by a chart of accounts (TAS).
Links between nodes are either contained in the data record for each grant block (e.g., all the levels of address are specified) or are implied by an external classification system (e.g., the program code is part of the CFDA and nestled in tree structures there).
The grant block record laid out in the initial guidance (see Figure 2) means, for example, we can follow a grant to a zip code, which puts this chunk of money in the context of a nested set of legal and political jurisdictions capped by states. Simultaneously, the same record can be represented in a variety of other classification systems, such as one organized by purpose (CFDA, the Catalog of Federal Domestic Assistance) or by a chart of accounts (TAS).
Links between nodes are either contained in the data record for each grant block (e.g., all the levels of address are specified) or are implied by an external classification system (e.g., the program code is part of the CFDA and nestled in tree structures there).
What’s Missing? Responsibility
The OMB model for “following the money” is missing the organization itself, the codes that place the grant within a chart of responsibilities.
There are two pieces of data missing.
-
First is the position responsible for each grant; and
- Second, where appropriate, the name of the person serving in the authorizing position.
Thus, for completeness and transparency on the input side—not just the output side–two more fields (with both a unique ID and a name/title) need adding to the recovery data model (see Figure 3).
Making the Recovery Map
The network logic of responsibility-accountability—a directed link between two nodes—provides a simple way to connect the key elements associated with each chunk of money. To illustrate, we connected four trees into a prototype network map of the recovery spending, including the missing organizational hierarchy. Each chunk, each block grant, shows up in four ways: by granting organization, recipient of the grant, the program and purpose for which it is used, and the recipient location.
As recovery dollars flow, so will the data stream start, and the mapping will be easiest. We have created an OrgScope model to demonstrate what a cross-linked network model might look like using an actual grant program by the NIH for external research facilities and an illustrative grant to a hospital in our city of Newton, MA (see org chart strip in Figure 1 in part 1).
-
For the organizational side, we used our network map of the top levels of the US Government (click to see this map live). To the existing org chart, which included NIH, we added the string of positions ending in a specific (dummy) grant block to Newton-Wellesley Hospital;
- Program category and type is modeled with a simplified outline of the Catalog of Federal Domestic Assistance (see CFDA.gov);
- Recipient type is illustrated through general categories of sector and industry; and
- Place is represented by a location map of Massachusetts organized by county (Middlesex), city (Newton), and zip code (02462, Newton Lower Falls).
Relationship chains relating a grant to the four major spokes of the recovery-spending wheel can be seen in the Figure 4 overview and the strings of connection shown in Figure 5 (see the live OrgScope map of these relationships). Of these, three have well-established classification systems (zips, DUNS numbers, CFDA). For the fourth, the government chart of responsibility for the funds, our digital org chart is a start to locating the program within the government structure. However, the government’s map of itself could and should extend to reach at least the program, if not the program officer levels.
It would not be too difficult to create an open government network map. At the top of departments and agencies, a few thousand positions in total are detailed in org charts that can be easily connected together, as we have shown in our publicly available USGov OrgScope map of many of these positions. For the two million or so civil service jobs, much of the information needed to assemble them almost automatically into org charts is just sitting there in federal financial systems. OMB may have much of it. The rest is in the many government human resource and information technology systems. Hooking the whole thing together is mostly a one-time expense of time and IT resources, and a process can easily be set up to refresh the basic data.
Transparency of organizational structure enables people both inside and outside the government to understand better the bureaucratic whole. By connecting the many bits of information into a coherent, shareable picture of our common government, we develop trust, hold one another accountable, and think through solutions with many minds.
Figure 5: Illustrative Grant in Four Interrelated Contexts
