LaVonda Atkinson laughs bitterly, shakes her head, and grins. "Your article" — the article you're now reading — "is gonna get me fired."
It is, after all, dangerous to stand in the path of a train. That's especially so in the case of Muni's long-gestating Central Subway project: a $1.6 billion, 1.7-mile extension to Chinatown and eventually, if vocal civic boosters get their way, along to Fisherman's Wharf for millions, if not billions, more. The Central Subway is a transit line. But it's also a monument to the coercive role of politics in the realm of transit planning. It's a pet project of the city's most powerful influence-peddlers, virtually every electable politician, and our federal representatives in Washington, D.C., disseminating largesse back to the home front. Atkinson is the project's cost engineer. It's her job to ensure the numbers add up.
Within three weeks of taking the job in July, Atkinson dialed the city's whistle-blower hotline, alleging large-scale accounting discrepancies and unethical behavior. She subsequently sent clandestine messages to the project's federal overseers warning of the same.
You're not going to believe this, but no one cared to join her in front of a train traveling upon tracks slicked with the highest grade of political grease.
And so, last month, she presented SF Weekly with reams of oft-conflicting Central Subway budgets and a trove of internal documents, serving as our Virgil in a descent into the underworld of Muni finances. From month to month, she demonstrates, budgets and expenditures attached to long-completed activities change, or line-items are added and drop out altogether. Budgets have been reverse-engineered such that — conveniently — they now match what was actually expended.
In the real world, work is undertaken and money is spent. But, on a spreadsheet, everything is malleable. As such, a $11.1 million project deficit printed in a December report blossomed into a $450,000 surplus in January and a $1.1 million bulge in February.
These numbers, incidentally, pertain to a phase of the project completed in July. So, that's a pretty neat trick.
Atkinson, the project's cost engineer — the person who ought to know — has no idea what the true budget and expenditures were for any of the hundreds of individual line-items at any given time. These numbers are altered, from report to report, even on activities that wrapped up years ago.
The budget for this massive transit project, she says, is nothing more than a vast game of three-card monte.
While Muni keeps tabs on how much money has been spent, it's far less certain just what that money was spent on. When it's unclear how much was spent on an activity, what was supposed to be spent, and how much work that spending bought, it becomes impossible to answer the overriding question affecting every San Francisco transit-rider and taxpayer: Will this project come in on budget? Or will the city, its transit agency, and its citizens have to cough up millions to fund a looming train wreck?
Aggrieved, would-be Muni passengers are familiar with the agency's ability to transcend the bounds of time and space. Buses and trains' estimated arrival times leap forward and backward — or drop, altogether, off the face of the earth.
Muni can do the same thing with money.
A March 19 email Atkinson sent to fellow members of the Central Subway project controls team nonchalantly reveals a "neg 40M delta" between costs reported to the city and to the Federal Transit Administration. In other words, Muni told the city controller it spent $109.5 million on preliminary engineering for the project — but reported to the FTA that it only spent $69.5 million on that. The $40 million difference was applied toward the project's construction phase. This, Atkinson says, creates the appearance of keeping a svelte budget during the prior phase and having cash on hand during the current one.
But without scouring each and every line-item, Atkinson says it's impossible to know which figure, if either, is truly accurate (in any event, she claims she's been told to stop reporting on this discrepancy).
So, Atkinson, the project's cost engineer — who, again, ought to know — has no idea if Muni's desired appearance matches its reality. Reality is a challenging concept: The agency, she says, has devised a budgetary wormhole to, at least on paper, beam funds backward and forward through time.
Every Central Subway line-item is assigned a "charge code." As far as the controller is concerned, you can tell when the money was spent based on the second of its 17 digits — a "1," "2," or "3" indicates the phase. Many of the activities, naturally, cut across phases. But Muni maintains just the one charge code, beaming the sums through time: "If it appears like you're having a large variance between your plan and your cost," Atkinson says, "You can change how much cost shows up when."
Wouldn't life be easier if you had a time machine? If, years ago, you overspent your budget, you could go back in time and simply change your budget to match what the costs turned out to be. But, Atkinson notes, there's a rub: Muni's time machine is a mere spreadsheet. You can change the numbers on a spreadsheet and make it appear budgets and costs were an exact match. In reality, though, what's done is done.
An overrun delayed is not an overrun denied. The reckoning will come.
It turns out there's a name for this practice: Transit expert and accountant Tom Rubin dubs it "the William Tell school of budgeting." Rubin, the former CFO of AC Transit and the Southern California Rapid Transit District and a longtime Central Subway skeptic, scoured documents supplied to SF Weekly by Atkinson and came away with the following assessment of Muni's financial planning for the massive project: "They fire the arrow at the barn and then draw the bull's-eye around where it hits."
While Muni officials tell SF Weekly that shuttling money between phases was an error, long since corrected, and limited to "earlier reports," documents provided by Atkinson indicate otherwise. Within the January 2014 "Detailed Monthly Expenditure Report," she points to the reassignment of $26.8 million worth of costs within the project's initial phase — which ended in January 2010. The cost engineer says she was tasked to fire up the time machine and move money around among long-completed activities in order to mitigate overruns, with whatever couldn't be swept under the rug beamed to a future phase of the project. That would be a problem for another day.