affected area and facilitate residential development in outlying areas. Regarding strategies that businesses in affected areas might take to minimize the impact on their employees, they could shift hours of operation, encourage telecommuting, reimburse workers who pay tolls, or relocate some or all of their operations to less congested areas. The complexity of the potential responses to congestion pricing makes it all the more difficult to predict its effect on urban form.

As an alternative to affecting land use patterns, congestion pricing could affect the balance between the location of jobs and residences. In other words, congestion pricing could induce workers to move closer to their work sites. Such shifts could substantially reduce the total amount of travel required for work trips. Commuting times that exist within the current distribution of jobs and residences in the Los Angeles metropolitan area are two-thirds longer than what would be required if workers were located in neighborhoods that minimized their commutes (Small and Song 1992). Whether congestion pricing fees would induce substantial changes in residential location, however, would depend on the strength of other influences on residential location choice. The extra commuting that occurs already may be explained by preferences for neighborhood amenities such as schools or low crime rates, the difficulty of minimizing commutes for both workers in dual-worker households, and other influences, such as racial discrimination (Giuliano and Small 1993).

The experience with pricing transportation in various modes provides a firm basis for predicting that congestion pricing would change traveler behavior and reduce congestion. The full range of potential behavioral responses is more complex than can be modeled with precision, but plausible estimates of the average motorist's response to price increases can be made. Congestion fees would cause some motorists to shift away from driving in the peak period. Estimates for the Bay Area and the Los Angeles area suggest that fees averaging $2.00 to $3.00/day on all thoroughfares in these regions would reduce peak-period travel by 10 to 15 percent and save average commuters 10 to 15 min per round trip.

Congestion pricing would have valuable systemwide effects. It would increase the efficiency with which existing assets are used by encouraging trip making in off-peak periods, ridesharing, and use of transit. If adopted