Although streetcars were at one point a significant mode of urban transportation in American cities, by the mid 20th C. they had fallen increasingly out of use and been displaced by busses and especially by private cars. The past couple of decades have seen a resurgence in consideration of streetcars by transportation planners, transit operators, and significantly by tax-paying voters, and the small handful of streetcar systems in the US surviving from that earlier epoch has begun to grow again.

As part of a team at the University of Utah’s Metropolitan Research Center I was involved in a study funded by the US Department of Housing and Urban Development, Utah Transit Authority, and the Wasatch Front Regional Council (WFRC), the metropolitan planning authority serving Salt Lake County and the northern portion of the greater Salt Lake City metro region. The research studied the impacts of new streetcar lines in Portland OR, Seattle WA, Salt Lake City UT, and New Orleans LA on economics, equity, and quality of life within the walksheds of new streetcar stops.

My most significant contribution to this project consisted of conceiving, automating, and executing a method for selecting “control” locations in each metro area that could be used as comparison sites. This allows us to measure changes in various economic and quality of life metrics over the time span before to after addition of new streetcar service and strengthens the soundness of our imputation of any observed effects to the “treatment” of adding a streetcar stop.

The process of selecting comparable sites consisted of two steps: first, creating a large set of candidate points, and second, assessing each candidate point for similarity with the actual streetcar locations. For the first step, we assume that bus service follows corridors similar to those considered by planners when designing streetcar alignments, and we create a large number (n=1000 in our work to date) of points randomly distributed along this network of bus routes. The second step uses census measures of population (POP), employment (EMP), housing units (HUN), households (HSH), and median HH income (INC), at the Census block group level from our identified “before” treatment year. These variables were chosen because they are available at annual intervals going back to the earliest year considered in our study (2003; the Seattle line); the block group is dictated as the spatial unit of reporting as this is the finest unit for which the Census Bureau reports the INC variable.

For each treatment point and candidate control point (i.e. selected streetcar location and randomly generated location on the bus network, respectively) we estimate values for each of the five variables for the area within one-quarter mile of the point, allocating block- group level measures proportionately to the fraction of the block group falling within this area. The quarter-mile buffer approximates the operational definition of a station-area walk-shed used in the qualitative portion of this study. The distributions of the five measures are checked for normality and, if necessary, transformed to normalize them; individual measures are then standardized (to z-scores). Taking each treatment point individually, we compute the distance in the standardized 5-dimensional space between the treatment point and each candidate control point.This distance is proportionate to dissimilarity,
and we select the ten candidate controls with the smallest distances as the control points for each treatment point.

Hinners, S., & Larice, M. (2016). Assessing the impacts of streetcars on economics, equity, and quality of life. Accessible here on the Metropolitan Research Center website.

Findings from this work with research collaborators Sarah Hinners and Chris Nelson analyzing these streetcar “treatment” and comparable “control” points has also been published in the peer-reviewed Transportation Research Record:

This study reports economic development outcomes—defined as change in employment—for areas within 0.25 mi of three streetcar stations along four lines in each of four cities: Portland OR; Salt Lake City UT; Seattle WA; and New Orleans LA. Employment 3 years prior to the start of the streetcar construction was compared with employment within the three streetcar station areas in 2013. Before–after comparisons were also applied to 10 bus stops within each urbanized area selected as “controls” because they were quantitatively comparable to each streetcar station area studied. Shift-share analysis is used to compare each streetcar and control station area’s employment trends with respect to central county economic performance. Portland’s system showed strong economic development consistency followed by New Orleans. The Salt Lake City and Seattle lines, however, did not perform as consistently. These results indicate that although streetcar investment may support economic development, it is not alone a driver of employment growth.

Hinners, S. J., Nelson, A. C., & Buchert, M. (2018). Streetcars and Economic Development: Do Streetcars Stimulate Employment Growth? Transportation Research Record. https://doi.org/10.1177/0361198118790096