Ten Thousand Hours of Privatization, Part Two
Weeks 10-13 of the Texas Voter Database Project
By Greg Moses
During weeks ten to thirteen (January 2005) of the Texas Voter Database Project we can see two significant transformations of the power matrix between the state and its private contractors. From one side, private contractors IBM and Hart InterCivic intensify their process of extracting from state employees their practical knowledge of election management, the better to privatize that knowledge into a commercial product.
From the other side, we see the state project manager imposing on Team IBM a meticulous process of plans and reports that discipline the private contractors into more transparent structures of information and accountability. The famous paperwork of state bureaucracy is force fed.
Also, in the process of writing up this report on the January activities, The Texas Civil Rights Review is beginning to feel the difference between reporting on state agencies and private companies.
For one thing, part of the reason that state agencies exist is to be involved in public accountability and criticism. We have logged a dozen hours this week viewing documents at the capitol. If the work resulting from docs supplied by the Secretary of State were to criticize the SOS, there might be consequences, but the game would be well known, and there would be no stock price to defend, no truly colossal sum of money at stake.
On the other hand, we notice that we do not spend any time at the headquarters of IBM or Hart InterCivic going through their documents. And if the results of our public research were to be perceived as critical of a global computer consulting firm, there might also be consequences, but the consequences would be of a different king (quite a nice typo I think). One effect of privatization therefore is to shift significant activities into realms where the games of public accountability and criticism meet new constraints, where also the consequences of debate are tossed into a context of stock prices and product sales. And when that happens, well, why do you think the business press is ever so cheery in comparison to the political press?
So if you are among the news consumers today who feel that even the political press has grown too cheery lately, perhaps that is just one more symptom of the shift we are all experiencing from a public to a privatized world order.
* * *
During the second week of January (week ten) the Colorado branch of Hart InterCivic invites a visit from two top-level administrators at the Elections Division of the SOS. It is difficult to imagine that Hart could do otherwise.
The makers of proprietary election equipment could hardly develop their wares with confidence if they did not know first what the public servants know about the practicalities of election management. The Colorado branch of Hart InterCivic first developed the eSlate voting system that we use in Travis County. Now they are developing a voter registration database, election management system, and jury wheel, all to be delivered to the state as a private product more technically known as COTS or Commercial Off-The-Shelf.
Although Hart is lead subcontractor for the project because of its ability to deliver an election management COTS, the product is not really ready for delivery yet. First, it has to be developed. As the visit to Colorado shows, the state is an active partner in helping Hart to develop the very same COTS that the state will license from Hart at a cost of $4 million, as soon as it is actually delivered.
In five years time the state will be able to buy back that COTS at “fair market value.” But first, as we say, the COTS must be developed and for this, forgive the repetition, public servants have to be invited to Colorado to meet with private contractors. And among those private contractors are private Subject Matter Experts who will later charge the state a hundred or two hundred bucks per hour to help solve any critical problems that might arise. Perhaps the reader is aware of famous economic principles of efficiency exemplified in this process. No doubt the principles are well known in business schools.
Also during week ten, Team IBM submits an invoice for $28,000 to cover its December work and a bill for $5,853 in travel and living expenses. A steering committee for the Secretary of State’s office (SOS) meets on Jan. 20.
During week eleven, billable hours are up to 4,230 and one more deliverable is delivered, but fifteen items “expected to be delivered” by Team IBM do not get delivered. We are beginning to understand why the state’s project manager would look back on this as a “poor start.”
Some of the early difficulties in getting started derive from the state’s failure to purchase equipment on schedule. The contract plan called for the state to purchase a million dollars in hardware and software (on top of the $12 million to contractors IBM and Hart InterCivic). But in a snafu officially known as Project Issue Number 001, it turns out the state couldn’t purchase computer equipment in such a straightforward way.
The problem with major computing tasks in Texas state government is that they are supposed to be consolidated into a San Angelo facility managed by Northrup Grumman. So the state project manager puts in a change order to move the project equipment to San Angelo as required by state law. It takes a couple of months to finally decide that the project will be kept in Austin as planned, but meanwhile Team IBM is having difficulty understanding the status of purchasing orders for hardware.
Back during week six the IBM project manager reported hearing from one person that the equipment had been ordered, while another person told her no, it had not. What she finds out during week seven is that the order cannot be placed for Austin equipment until the purchase is authorized by means of a special waiver from the Legislative Budget Board, since the hardware will not be placed in San Angelo.
Meanwhile, thanks to that meeting in Colorado, Team IBM reports that critical requirements have been collected for two crucial pieces of the election management puzzle: Election Night Reporting (ENR) and Ballot Definition (BD).
The Team IBM project manager is having troubles of her own keeping up with meticulous administrative details demanded by the state’s project manager. As she turns in a new, revised work plan that is “deliverable driven rather than task driven” she makes a formal complaint for the record. “Increased administrative tasks have caused deliverables and work papers to slide.” It is pleasing to see how she appropriates the new language of deliverables in order to articulate her complaint. She logs her complaint as Project Issue Number 002, assigns it a priority level ‘H’ for high, and places it on an Issues List to be either analyzed, brought up for decision, or resolved during future reporting periods. In fact, whether she knows it or not, her days as IBM project manager will soon end.
In the hardware/software area, a “development environment” has been set up where Team IBM can share project files with the state, but there is still no election system to test yet, because first of all the hardware hasn’t arrived, and second of all the VR (or voter registration software) is still in demo stage at Hart.
During week twelve, the Team IBM administrator uses a brand new status report format (version four) to say that she has helped to produce two new plan docs, two strategy docs, and “revised project work plan number 80.” And on Jan. 28, the project convenes a Focus Group of 32 people from 15 counties to share progress to date.
As January changes to February, ten more plans are delivered during week thirteen while intensive meetings continue in Colorado. As Team IBM also convenes meetings with the state IT staff to discuss ENR and BD ar
chitecture, Hart says that its Voter Registration build is 90 percent complete, awaiting crucial review by one state expert who happens to be not available at the moment. A third “project issue” is added: when will the scope of ENR and BD be defined?
And finally, in the state’s report for week thirteen, original completion dates are x-ed out and moved back. In week thirteen the project is officially behind schedule.
To be continued in series.