Stacking the deck, packing in development

FOR YEARS NOW, residents of the sprawling suburbs on the Eastside's eastern edge have wondered what the hell is going on. The area known as the East Lake Sammamish Plateau was already choked with traffic—but gargantuan new housing developments kept winning county approval. Supposedly, the state's Growth Management Act lets cities and counties permit only as much development as their roads will support. But it was evident that the Plateau's few roads could barely handle present loads, let alone the thousands of minivans and Mercedes to come.

Now the mystery may be answered. For the first time, an outside expert has gotten to peek inside the computerized traffic-forecasting system the county uses to decide whether to approve new housing developments. He found the county's software to be, in essence, rigged in favor of developers—designed to consistently underestimate the traffic impacts of new housing projects and grease the skids to approval.

This finding has set off alarms at the county, whose transportation officials are this week responding to the expert's charges in a case before a county hearing examiner. "This is very big," says King County Council member Brian Derdowski, a longtime growth opponent who represents the Plateau. Should the examiner rule against the Department of Transportation (in a decision expected sometime after Labor Day), the entire county system for approving developments could be brought down—not just for future Estates and Chateaux, but for some that are already under way.

The case arose over a proposed development called the Greens at Beaver Crest, in the heart of the Plateau. Just two years ago, Scott Hamilton had moved to the area from suburban Dallas in search of a better lifestyle, and bought a home on a wooded dead-end street. But the 254-unit Greens, together with 1,000 other units the county had approved nearby, promised to send some 10,000 vehicles a day past his front door. In January, Hamilton appealed the Greens' green light. "I didn't have the foggiest idea what I was getting into," he says. "I was just hoping for some mitigation."

KING COUNTY government regulates all "unincorporated" land in the county—areas like the Sammamish Plateau that aren't part of any city. To comply with the Growth Management Act, every new development must pass a "concurrency" test. The county's computer model estimates how many cars a development will disgorge, and onto which roads. It looks at those roads' expected future capacity. And it translates these forecasts into a score. If a project's traffic score comes in below a cutoff point, it receives its "certificate of concurrency." If its traffic score is too high, it is rejected and has to slim down. The idea, central to growth management, is that development should occur only in tandem with transportation infrastructure.

This may sound simple, but the evaluation is fantastically complex, involving hundreds of computations and hazy predictions as to how many people all over the region will drive where on which routes. Though the concurrency test has become the main gatekeeper for county growth since its adoption in 1995, it's remained completely inscrutable and unaccountable to all but a few traffic programmers.

Until now. In his appeal, Hamilton requested a copy of the traffic program. The county had rebuffed all such requests before, but the hearing examiner affirmed Hamilton's motion, and the county eventually complied. Hamilton then enlisted Bellevue traffic consultant Joseph Savage to scrutinize the program.

Savage is no populist tree hugger. His consulting firm, KJS Associates, does most of its work for private developers (including, in the late '80s, the giant Plateau development Klahanie). "Developers who meet zoning requirements should rarely be denied," he says.

But in taking apart the county's software, Savage found what he calls "a pattern of errors" with one common characteristic: "In every case, the errors lead to a lower [traffic] score." He contends the county model will "consistently underestimate" the traffic impact of developments.

Savage points to many actual lapses, including the model's failure to consider explosive growth in the cities (e.g., Redmond, Issaquah) that surround the county land. But the "biggest, most obvious" factor, he says, is a piece of code in the program that automatically lowers the score any time a development is projected to produce too much traffic for a given road: "They artificially deflate the score without any particular justification."