The worst part is there is no defense against it as of yet. Traffic planning as it is done today is accepted as omniscient and therefore it is omnipotent. Unfortunately, it is badly flawed in many ways, not least of which a fundamental misguided approach and misunderstanding (perhaps willfully) of how cities 1) work and 2) ought to work in order to achieve 1) order 2) intelligence and 3) as an interdependent product of the first two, a multiplicity and availability of choice.
You can find those here:
The 4 blind spots in transportation planning.
Where does the Traffic go?
How Less Capacity Makes Us More Mobile.
Don't believe me? Good. You should be skeptical of all information coming at you. Including those coming from transportation planners telling you they will "improve" your road. Keep in mind the difference in metrics. Improve to them means improved level of service. This means faster flow, increased severity of collisions, less pedestrians, and ultimately less safety, if not complete abandonment of place by way of disinvestment. Sounds rad, huh?
You on the other hand as a resident, somewhere, meaning you have a stake in your neighborhood. You want it to be livable and functional. It should be walkable and safe. If there are no pedestrians, there are no pedestrians to be run over. This is the mindset of the common traffic engineer. To paraphrase Ronald Reagan (of all people), "I'm from the DOT, I'm here to help."
Fortunately for the skeptical mind, there are ever more rigorous academic studies pointing out what you've already intuited: that transportation planning is fundamentally flawed. When you back down because they have fancy math and metrics, you can be increasingly armed with studies and similar data to shoot down their findings, such as this from the Aalborg University in Denmark. A summary:
Although the phenomenon of induced traffic has been theorized for more than 60 years and is now widely accepted among transport researchers, the traffic-generating effects of road capacity expansion are still often neglected in transport modelling. Such omission can lead to serious bias in the assessments of environmental impacts as well as the economic viability of proposed road projects, especially in situations where there is a latent demand for more road capacity.
Reading through the study, the researchers ran two scenarios, one (A) with attempting to quantify induced demand based on findings from other international studies and one (B) running the projections the typical way. Perhaps the most interesting finding is related to the return on investment when calculating for time saved on a per currency (in this case DKK) basis. They found that a Braess paradox occurs without significant capacity increase. Meaning, a slight increase in congested areas yields even slower traffic through the new demand created.
Furthermore, if a great amount of new capacity is created the costs still outweigh the benefits. It's a lose-lose either way despite what the flawed existing model suggests. Either spend too much on infrastructure over-widening roads at very high costs of construction and to property or add capacity at a minimal level worsening the situation, including environmental effects from more cars, idling, and pollution.