Our current alarm about traffic congestion stems in large part from perception of trends: thirty years ago traffic flowed smoothly; today it crawls. If this trend continues, congestion will become gridlock. These perceptions lead to statements such as: "There is no point to building highways, new lanes fill up as soon as they are opened."
I present evidence to show that such trend-based thinking is wrong because it ignores structural shifts in the demographics of auto ownership and use. At this time, auto ownership is effectively saturated: we are very close to the point where all the potential drivers have auto access. The ratio of autos per driver can continue to grow, but since it is only possible to drive one vehicle at a time, the growth rate of auto-use must decline to about the rate of population growth -- a rate which is 2.9 times lower than the rate we have experienced in the period since 1960.
Thus, fatalistic prophesies about future gridlock have overstated the potential growth of demand for auto travel. That growth has already declined and it should level off to a rate which is only one-third as large as we are used to. This is a manageable rate, planning is possible. And, specifically, it is appropriate to think about building new roads to solve our deficiency of highway capacity.