Introduction

Traffic engineers often compare traffic flow to a fluid, but in congested situations it is better described as a gas, which expands to fill available space. Transport improvements that reduce user costs (including vehicle operating costs and travel time) tend to attract trips from other routes, times and modes, and increase total travel. This is called generated traffic or induced travel. Economic analysis requires that all incremental (net) benefits and costs be including in the evaluation of policies, programs and projects. Incremental costs and benefits to existing trips are relatively easy to determine, but special consideration is needed to determine net benefits and costs of generated traffic.

Generated traffic has three implications
for transportation decision making.

1. First, generated traffic provides relatively less benefits than existing trips, since it consists of lower value trips that consumers are most willing to forego due to congestion delays.

2. Second, generated traffic erodes much of the predicted congestion reduction. Studies have found that 40-90% of added highway capacity is filled with generated traffic, and in some cases congestion costs actually increase. One researcher concludes,

"At the country level we find that a 1.0 percent increase in lane-miles soon induces an immediate 0.2 percent increase in traffic, building to a 0.6 percent increase within two years after the lane-miles are added. At the metropolitan level, the immediate effect is also about 0.2 percent, building to a 0.9 percent increase after four years. Therefore, it appears that adding road capacity does little to decrease congestion because of the substantial induced traffic."

3. The third implication is that generated travel increases external costs such as accidents, congestion on other roads (for example, increasing freeway capacity oiljkincreases traffic volumes and therefore congestion on surface streets), parking demand (and parking subsidies), noise, air pollution and energy consumption. These external costs are significant, particularly under urban peak-period conditions.

While users' marginal benefits exceed their marginal costs (if not, users would not take the additional trips), these benefits do not necessarily exceed total incremental costs. Generated traffic, therefore, may create more costs than benefits.

Failing to incorporate generated traffic into transport modeling and planning, underestimating generated traffic, or ignoring any of these three factors can significantly affect transport decision making. A report by the UK Standing Advisory Committee on Trunk Road Assessment concludes, "... the economic value of a [road] scheme can be overestimated by the omission of even a small amount of induced traffic. We consider this matter of profound importance to the value-for-money assessment of the road programme."

Models that fail to consider generated traffic were found by one study to overvalue roadway capacity expansion benefits by 50% or more. One researcher states,
"...at moderate levels of congestion and elasticity values, the fixed matrix measure of benefits [which ignores generated traffic] exceeds the variable demand measure [which incorporates generated traffic] by 20-50 per cent. An important finding is that quite small absolute changes in traffic volumes have a significant impact on the benefit measures. Of course, the proportional effect on scheme Net Present Value will be greater still."

Failing to incorporate feedback (congestion impacts on future travel) into traffic modeling can significantly overpredict traffic congestion problems. One study found that incorporating feedback when modeling a congested road network increased predicted traffic speeds by more than 20%, and reduced predicted VMT by more than 10%. Another study found that the ranking of preferred projects changed significantly when generated traffic feedback is incorporated into project assessment analysis. Specifically, capacity expansion options provide less congestion reduction benefit and increase air emissions, while demand management and No Build options offer greater benefits.

Traffic models used in some large urban areas incorporate generated traffic feedback, but traffic models used in medium and small communities usually do not, and generated traffic costs are seldom incorporated in the economic analysis of specific projects. This omission is often justified with the excuse that tools do not exist to predict how much traffic will be generated or resulting net costs. These excuses are no longer valid.