Safety Evaluation of Continuous Green T Intersections

CHAPTER 8. ECONOMIC ANALYSIS

A B/C analysis compared the safety benefits with the construction costs of a CGT relative to a conventional signalized three-leg intersection. This chapter describes the assumptions used in the analysis, describes the differences in the construction costs between the CGT and the conventional signalized three-leg intersection, derives the safety benefits associated with the CGT, and computes the B/C ratio for the CGT relative to a conventional signalized three-leg intersection.

ASSUMPTIONS

Because this study was unable to use an observational before-after study methodology, the B/C analysis presented in this report compared two different intersection forms (CGT versus conventional three-leg signalized intersection). To complete the B/C analysis, the following assumptions were made:

• The median width was as wide as the left-turn lane adjacent to the continuous flow lane on the major street.
  
  
• The median was unpaved.
  
  
• There were shoulders on both the major and minor streets.
  
  
• There was no fourth leg at the intersection.
  
  
• There were fewer than five through lanes.
  
  
• The comparison intersection was a signalized T intersection.
  
  
• The pavement design life was 20 years.
  
  
• The average traffic volume for the comparison group was constant over the 20-year period (to be used for predictions).
  
  
• All safety benefits were derived using the South Carolina data, so the Florida indicator variable in table 13 was set equal to zero. It should be noted that the Florida indicator variable was negative, so the B/C ratios for Florida were larger than those computed using the South Carolina data.
  
  
• There were no maintenance costs because the project design life was equivalent to the pavement design life.
  
  
• The existing traffic signals could be used for the CGT intersection.
  
  
• The only cost associated with the treatment was the additional pavement for the acceleration lane, as shown in figure 20.

Figure 20. Schematic. Traditional and CGT intersections.

The CMFs used for the evaluation were those estimated using the propensity scores-potential outcomes framework (genetic matching results (table 13)). The treatment cost was dependent on the posted speed limit. A minimum (35 mi/h on the major road) and maximum (55 mi/h on the major road) cost for the treatment are estimated in figure 21 and figure 22.

For the new pavement, the low (posted speed = 35 mi/h) value required is as follows:

• 210 ft to the beginning of the taper at 12 ft wide = 210 x 12 = 2,520 ft².
• 125-ft taper with an average of 6 ft wide = 125 x 6 = 750 ft².

Figure 21. Equation. Total pavement required for low posted speed.

For the new pavement, the high (posted speed = 55 mi/h) value required is as follows:

• 435 ft to the beginning of the taper at 12 ft wide = 435 x 12 = 5,220 ft².
• 540-ft taper with an average of 6 ft wide = 540 x 6 = 3,240 ft².

Figure 22. Equation. Total pavement required for high posted speed.

The cost for asphalt pavement used for the analysis was $28/yd². The cost for concrete pavement was $70/yd². Thus, the cost for 35 mi/h was $10,173.33 for asphalt and $25,433.33 for concrete. The cost for 55 mi/h was $26,320 for asphalt and $65,800 for concrete.⁽³⁵⁾

The number of crashes (total, fatal and injury, and property damage only (PDO)) with and without the CGT, using the average AADTs from the comparison group, were predicted for the 20-year service life. For example, the total expected number of crashes for the untreated intersections were computed using the equation in figure 23, assuming that the posted speed limit on the major road was 35 mi/h (indicator variable for Thru_Spd was set equal to zero because it is the baseline value), minor (intersecting) roadway was 35 mi/h (indicator variable for Int_Spd_35 in table 13 was 0.494), the site was a comparison site (treated variable was set equal to zero), and the intersection was located in South Carolina (Florida indicator was zero).

Figure 23. Equation. Total number of expected crashes for the untreated intersections.

N_total = Total number of expected crashes.
e = The exponential function.
through = The subscript related to through street traffic volume (veh/day).
intersecting = The subscript related to the intersecting road traffic volume (veh/day).

The descriptive statistics for the South Carolina comparison group are shown in table 8 and table 9, and the average through and intersecting roadway AADT volumes are 22,452 and 8,452 vehicles per day, respectively. Inputting these values produced the expected number of total crashes per year, as seen in figure 24.

Figure 24. Equation. Total number of expected crashes per year for the untreated intersections.

As such, the expected annual total crash frequency for the South Carolina comparison group sites was 7.20 crashes per year, as shown in table 16. Multiplying the annual crash frequency by 20years produces 144 crashes. The number of property damage only (PDO) crashes was estimated by subtracting the number of fatal and injury crashes from the total crashes. The treated crash frequency predictions were derived by applying the CMFs shown in table 13. The total treated crash frequency estimates were derived by multiplying 144 crashes (untreated crashes) times the CMF for total crashes. This resulted in 144 times 0.958, which equals 137.95 crashes over a 20-year period. All of the predicted crash frequency estimates for the untreated and treated intersections are shown in table 16.

The comprehensive crash costs used for this analysis were derived using 2001 dollar values from Council et al.⁽³⁸⁾ As suggested by the authors, the crash cost values were multiplied by the ratio of the Consumer Price Index for 2001 and 2014. This ratio was 2.425. The 2001 comprehensive crash costs were $129,418 for fatal and injury crashes and $10,249 for PDO crashes on roads with posted speed limits below 50 mi/h. The 2001 comprehensive crash costs were $146,281 for fatal and injury crashes and $4,015 for PDO crashes on roads with posted speed limits equal to or above 50 mi/h. This produces crash cost savings of $1,536,250 for the 35 mi/h posted speed and $2,427,752 for the 55 mi/h posted speed limit for the 20-year project life. The annual benefits (from crash costs) were $76,813 for the 35 mi/h posted speed limit major roads and $121,388 for the 55 mi/h posted speed limit major roads. Thus, the B/C ratio, by pavement type and posted speed limit, were estimated and are provided in table 17.

The annual costs (based on the initial paving costs and no maintenance over the 20-year project life), discounted at 7 percent over the 20-year project life, were $956.30 for asphalt and $2390.70 for concrete pavements at 35 mi/h intersections and $2474.10 for asphalt and $6185.20 for concrete at 55 mi/h intersections, respectively.

Sensitivity Analysis

In order to test the sensitivity of the B/C ratios to variability in the safety benefits, the upper and lower bound of the 95-percent CI of the CMF estimate for total crashes in table 13 was applied to the safety benefit estimates shown in table 17. This produced B/C ratios that ranged from 62.0 to 95.5 for the 35 mi/h posted speed limit on asphalt pavements and from 37.9 to 58.4 for the 55mi/h posted speed limit on asphalt pavement. The B/C ratio ranged from 24.8 to 38.2 for 35mi/h posted speed limits on concrete pavements and from 15.1 to 23.3 for 55mi/h posted speed limits on concrete pavements.

Further sensitivity analysis was done to determine the construction costs that would still achieve a B/C ratio of 2.0 (lower bound) for the 35 and 55 mi/h posted speed limits. (The crash costs were equal for asphalt and concrete pavements.) For the 35 mi/h posted speed limit, a B/C ratio of 2.0 could be achieved with annual construction costs up to $38,407. For a 55 mi/h posted speed limit, annual construction costs up to $60,694 produce a B/C ratio up to 2.0.