Practices for Establishing Contract Completion Dates for Highway Projects (2017)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

12 GENERAL CHARACTERISTICS OF THE CURRENT TOOLS To collect state-of-the-practice information, a web-based survey was distributed to the voting members on the AASHTO Subcommittee on Design. Respondents from 41 state departments of transportation (DOTs) provided information on the contract duration determination procedure. Among the 41 respondents, 28 (68.3%) indicated that a formal, documented procedure for determin- ing contract duration was in use at their agencies, whereas the other 13 (31.7%) indicated they did not have a documented procedure. Most of the current contract duration estimating tools are customized to suit the agencies’ need. Only a few are purchased off-the-shelf (Figure 5). Respondents commented that off-the-shelf CPM scheduling software is used in determining contract durations. Among the 28 states that have formal contract duration estimating procedures, 12 (42.9%) responded that the design division handled contract duration estimation, whereas seven (25.0%) answered that the task was performed by the construction division. Figure 6 shows the distribution of respondents’ answers with a pie chart. Among the eight (8) respondents that answered “Other-Please Specify,” one (1) indicated that the agency had a designated estimating office, two (2) that the contract office within their respective agency estimated contract duration, two (2) that estimating contract duration required joint effort by design and construction personnel, and two (2) that estimating contract duration was carried out in district (or local) offices. Survey responses show that state DOTs specify contract durations or dates in different ways (calen- dar days, working days, or completion date). About half of the state DOTs that responded to the survey primarily use a completion date to mark the end of a contract. Working day is the next commonly used method to specify contract durations. Only three (3) respondents indicated that their agencies primar- ily use calendar days to specify contract durations. Figure 7 shows the result of the survey responses as to how state DOTs specify contract duration and time. Each bar in Figure 7 represents a response. The vertical axis indicates the percentage of projects in which the agency specifies contract duration and time with each of the three methods. Figure 7 also shows that most state DOTs use one method to specify contract duration and date for projects. Very few states use a combination of the three methods. Eight (8) of the 28 respondents indicated that anticipated contract duration was considered when determining which method to use for specifying contract duration and date (e.g., for a project with a very short anticipated duration, the DOT may choose to set a completion date rather than establish a project duration in calendar days or working days). DETAILED DESCRIPTIONS OF THE CURRENT TOOLS Overall Description Survey participants were asked how they would describe the contract duration estimating tools at their agencies (Figure 8), and the most frequently selected answer was “simple.” Eighteen of the 21 respondents who provided answers to this question believed that the current contract dura- tion estimating tools at their agencies were simple. Eight (8) respondents reported that the tools at their agencies were sophisticated. Five (5) selected both “simple” and “sophisticated,” and two (2) chapter three RESULTS OF CONTRACT DURATION SURVEY

13 Purchased off the shelf 11% Custom developed 53% Purchased off the shelf and then customized 11% Other 25% FIGURE 5 Source of the current contract duration estimating tools. Design 42% Procurement 4% Construcon 25% Other - please specify 29% FIGURE 6 Division that performs contract duration estimates. Response ID from the Survey 43 54 62 37 49 50 39 22 25 38 46 26 27 56 32 24 53 28 36 55 30 31 57 65 47 35 64 34 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Calendar Day Working Day Comple on Date % o f C on tr ac t FIGURE 7 Methods used at state DOTs to specify contract durations and dates.

14 pointed out that different procedures might be used depending on the size and complexity of a proj- ect. Twelve (12) survey participants noted that some calculation by hand was required to estimate contract duration at their agencies, whereas nine reported that their procedures were automated (or program assisted). Twelve (12) participants believed the tools for estimating contract durations were effective; however, one participant believed the current tools were ineffective. Responses from survey participants also indicated that the ability to use the current contract duration estimat- ing tools depended on construction experience. Four (4) respondents noted that having the formal duration estimating tool in place helped reduce claims for additional time and that the tool was an asset to managing agency personnel. Guidance of the Tools The survey asked the participants what guidance contributed to the procedure used for estimating contract duration at their respective agencies, and the responses indicated that most tools currently in use by state DOTs were guided by manuals developed by internal personnel. Federal guidance and contractor input were the second and third most frequently selected options. A few states use tools developed by universities and other third-party organizations (Figure 9). 0 2 4 6 8 10 12 14 16 18 20 Number of Respondents Simple Sophiscated Calculated by hand Program Assisted/Automated Effecve Ineffecve Construcon Experience Dependent Labor Intensive Reduces Claims for Addional Time Is an Asset to Managing Agency Personnel 18 8 12 9 12 1 12 5 4 4 Note: Respondents were asked to check all that apply FIGURE 8 Descriptions of the current contract duration estimating tools. 1 2 3 4 4 6 22 University/Transportation Research Developed- Related to Other Agencies Federal Regulations Developed/Adapted by Third Party University/Transportation Research Developed- Specific to Your Agency Contractor Input Federal Guidance Developed/derived by Internal Personnel/Efforts Number of Respondents Note: Respondents were asked to check all that apply 0 5 10 15 20 25 FIGURE 9 Guidance contributing to the current contract duration estimating tools.

15 0 0 3 4 5 6 6 10 11 14 14 16 16 0 2 4 6 8 10 12 14 16 18 Cost/Value Loaded Methods Unsure Equations to Estimate Project Durations (potentially regression based) Critical Path Method-using static logic Intelligent Production Rates (e.g., adjusted for location, project scale) Generic Production Rates (rates from outside sources) Contractor Input Construction Personnel Verification Historic Project Production Rates (agency based and frequently updated) Static Production Rates (agency based, just not frequently updated) Construction Personnel Estimates Critical Path Method-using project specific logic Historic Project Durations Number of Respondents check all that apply Note: Respondents were asked to FIGURE 10 Attributes used by current contract duration estimating tools. Methods Used According to survey responses, CPM for scheduling is widely used by state DOTs for estimating con- tract durations. Contract durations are more frequently estimated using project-specific logic rather than pre-defined logic when using CPM. Figure 10 shows the frequencies of the methods selected by survey respondents. Many states use agency-specific production rates to estimate time required to complete each task within a project. A few use generic production rates (rates from estimating manuals, not agency- specific). Of the 22 states that use agency-specific production rates, 14 (63.6%) reported that their rates were not frequently updated, and only 11 (50%) that their rates were frequently updated. Several states have applied more advanced methods for estimating contract durations. Five (5) states indicated that they adjusted production rates for factors such as project size and location. Three (3) participants responded that they used regression-based equations. Criteria and Factors Considered Figure 11 shows the survey responses when participants were asked what criteria or factors were taken into consideration by the procedure currently in use at their agencies. The responses indicate that the most common items of consideration in estimating contract time include project phasing, maintenance of traffic, time of letting and seasonal limitations, environmental limitations, and other. Interestingly, the results indicate that while the time of year of the letting is considered, not much consideration is

16 given if or when lettings are delayed. This may indicate the need to emphasize contract time reviews and revisions when letting delays occur. Another pattern observed from the survey responses is that although many states reported using production rates to estimate contract durations as suggested by FHWA’s guidance, some factors that significantly impact potential contractors’ production rates (such as the use of multiple shifts, potential contractor’s equipment, and production capability and project size) are not often taken into consideration. In addition to the criteria and factors shown in Figure 11, follow-up interview with one of the respondents also revealed that there were overriding factors, such as promises to the public, which may dictate the starting date and completion date of a particular project. Special Conditions The survey questionnaire asked participants to select any special conditions considered when determin- ing contract dates and durations at their agencies. The questionnaire provided three options including 0 5 6 7 9 12 14 17 17 18 18 19 20 20 21 22 22 23 0 5 10 15 20 25 Market Conditions (e.g., Material Price Index) Economies of Scale in Bid Item Quantities for Production Rates Letting/Award Delays Require Maintenance of Access Potential Contractor Equipment & Production Capability Project Size Effect on Production Rates Multiple Shifts Weather Right-of-Way Limitations Utility Conflicts/Limitations Milestones Long-Lead Time Materials Seasonal Shutdowns and Seasonal Limitations Permit/Environmental Restrictions Community Factors (School Sessions, Fairs, Conventions, Events) Letting/Award Time of Year Maintenance of Traffic Project Phasing Number of Respondents Note: Respondents were asked to check all that apply FIGURE 11 Criteria and factors considered by current contract duration estimating methods.

17 “Lane Rental,” “Cost-Plus-Time Bidding,” and “Incentives/Disincentives,” and the most frequently selected option (23 of 28 respondents) was “Incentives/Disincentives” (Figure 12). Aside from the three options provided, one respondent answered that hourly liquidated damages for extended lane closures was also considered. Incentives and disincentives appear to have a discernible impact on contract duration. Of the 23 states that responded they used incentives and disincentives, 78% believed that using incentives and disincentives influenced the on-time completion of contracts. The remaining 5 (22%) reported they were unsure if such influence existed. Survey responses also revealed that in general state DOTs pay out incentives more often than they collect disincentives (Figure 13). Contractor Performance Measurement Survey responses show that not all state DOTs evaluate contractors’ schedule performance by compar- ing actual contract durations with the estimated durations. Among the 28 states that provided informa- tion on contractors’ duration performance measurement, the most commonly used method is tracking the percentage of projects that experienced duration overrun (Figure 14), which suggests that state DOTs are mostly concerned when projects are not completed on time, but do not spend as much effort to track the percentage of projects that are completed ahead of schedule. Responses also suggest that state DOTs are more concerned about the number of projects that require changes in their estimated contract times than the magnitude of duration change within one project. One survey participant also indicated that its agency tracks the amount of incentive and disincentive paid and collected to evaluate contractors’ schedule performance. N um be r of R es po nd en ts 0 5 10 15 20 25 13 19 23 Lane Rental Cost-Plus-Time (A+B) Bidding Incenves and Disincenves FIGURE 12 Special conditions considered by current contract duration estimating tools. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Never Rarely Rarely Some mes Some mes Oen Oen % o f R es po nd en ts Disincen vesIncen ves FIGURE 13 Frequency of contract incentives and disincentives.

18 Another survey participant stated that contractors’ schedule performance was evaluated by com- paring actual contract time against the original estimate; however, the evaluation results were not reported back to those who conducted the estimate. Therefore, schedule deviation from the original estimate was not used to evaluate the accuracy of the estimating procedure. EFFECTIVENESS OF THE CURRENT TOOLS The survey questionnaire asked participants if their agencies had performed any evaluation on the effec- tiveness of the current contract time estimating tools; eight of the 27 participants who provided infor- mation to this question (29.6%) answered “Yes.” The remaining 69.4% either indicated no evaluation on the effectiveness of the current tools had been performed or were unsure (Figure 15). As previously discussed, at some state DOTs, actual project durations are compared with estimated durations to evalu- ate contractors’ performance, but not necessarily used to measure the accuracy of the current estimating tools. One STA provided analysis results of their on-time performance for projects for three successive quarters that indicated an average on-time performance of 61% across the three quarters. To examine the accuracy of the current estimating procedures, the survey questionnaire asked participants to indicate, to the best of their knowledge, how accurate their estimates were compared with actual project durations. Thirty-three percent (33%) of the respondents (10 of 27) answered they were unsure (Figure 16). Of the 17 respondents who provided information on the accuracy of their procedures, nine indicated the average percentage difference between estimated and actual project durations was 1 3 6 9 11 0 2 4 6 8 10 12 Amount of Incentives/Disincentives Percent change within Individual Projects Percent of projects under set contract duration Percent of projects requiring a change to contract duration Percent of projects over set contract duration Number of Respondents Note: Respondents were asked to check all that apply FIGURE 14 Methods for measuring contractors’ schedule performance selected by survey participants. Yes 30% No 55% Unsure 15% FIGURE 15 Survey responses to whether the agencies evaluate the effectiveness of the current contract time estimating procedure.

19 between 0% and 25%, seven that the percentage difference was between 50% and 100%, and one that the percentage difference was between 25% and 50%. In general, the responses show that at many STAs the current estimating tools are not predicting contract times with acceptable accuracy. The accuracy of the contract time estimating tools was compared between the STAs that performed analyses on the tool’s effectiveness and the STAs that had not performed any analysis, with results showing that the former were generally more accurate (Figure 17), suggesting that performing analysis on the effectiveness of the current tools likely helped improve the tool’s accuracy. Of the eight (8) STAs that evaluated the effectiveness of their estimating procedures, four (4) indicated that the errors of their estimates were within ±25% and two (2) that their estimates were off on average by ±50%–100%. The reported accuracy by which divisions within the STAs performed contract time estimates was also compared. Results showed that at STAs where the construction division handles contract time determination the accuracy level is generally better than at STAs where other divisions perform the estimates (Figure 18). The finding is consistent with the point made by some of the survey respondents that the current tools are highly dependent on interpretation by experienced construction personnel. PROCEDURES FOR ALTERNATIVE PROJECT DELIVERY METHODS Design-bid-build is the traditional, and still overwhelmingly common, project delivery method used by STAs. However, in recent year some STAs have adopted other delivery methods, such as design- build and CM/GC. Figure 19 shows the number of STA survey participants that have developed +/-0-25% 33% +/-25-50% 4% +/-50-100% 26% Unsure 37% FIGURE 16 Survey responses on accuracy of the current tools. Note: Column widths are proportional to the numbers of responses in each group (“Yes”, “no” or "Unsure”). Block heights are proportional to the numbers of respondents who selected the corresponding accuracy level within each group. FIGURE 17 Accuracy of estimates by whether or not a DOT performs analysis on effectiveness of the estimating tool.

20 official procedures for the three delivery methods. Most of the participating states (27 of 28) have official procedures for design-bid-build projects, fewer than half have official procedures for design- build projects, and only two (2) have official procedures for CM/GC projects. There was an interest in finding out whether the procedures for determining contract time were the same for the three delivery methods. Table 2 summarizes the responses collected through the survey. Among the 26 respondents who provided answers to the corresponding question in the survey, 15 indicated they used the same procedure for all delivery methods, and nine that there were some differences among procedures for alternative delivery methods. For these nine respondents, the survey further asked about guidance and methods used and factors and conditions considered for each of the procedures for the three delivery methods (Questions 14, 16, 18, and 20 in Appendix A) in the hope of identifying the differences among procedures. However, few respondents were able to provide information on the procedures for design-build and CM/GC delivery methods (Appendix B, Questions 14, 16, 18, and 20; columns for design-bid-build methods received sufficient response rates, but columns for the alternative delivery methods received low response rates). One pattern the team was able to identify was that, for design-build contracts and CM/GC contracts, third parties and contractors contribute more when determining contract durations/dates. Note: Column widths are proportional to the numbers of responses in each group (“Construction”, “Design” or “Other”). Block heights are proportional to the numbers of respondents who selected the corresponding accuracy level within each group. FIGURE 18 Accuracy of estimates by which division estimates contract time. 27 13 2 Design-Bid-Build Design-Build Construcon Management/General Contractor FIGURE 19 Number of participating DOTs that have contract duration estimating procedures for different project delivery methods.

21 Low response rates for alternative delivery methods likely suggest that even at STAs where projects with alternative delivery methods are handled differently than traditional design-bid-build projects, the official documented contract times estimating procedures are still tailored to fit the design-bid- build delivery method. The procedures for alternative delivery methods may not be well-established or well-documented; therefore, it is difficult for respondents to specify exactly where the differences lie. According to the survey responses, several states are planning to improve or refine the contract time estimating procedures for alternative project delivery methods. Three (3) respondents reported that their agencies had plans to improve or refine the procedures for design-build projects and two (2) for CM/GC projects. IMPROVEMENT NEEDED At the end of the survey questionnaire, participants were asked to identify areas of improvement needed for the current contract time estimating procedures. Responses are displayed in Figure 20. Only two respondents answered that no improvement was needed. Seventeen (17) of the 27 respon- dents believed more feedback and communication between agency divisions was needed in order for the contract time estimating procedures to be more effective. Improved accuracy is the second most frequent concern among survey respondents. A few respondents pointed out that for the current tools to be more accurate, the agencies would need to update the production rates more frequently. Nine (9) respondents believed their current procedures needed to be improved to be able to adapt to multiple delivery methods. Increased automation and improved usability are also frequently selected by survey respondents. One respondent stated that agency staff lacked motivation to use the current estimating tool. Increased automation and improved usability may help increase the popularity of the estimating tools among agency staff. Response % Count Same procedure for all delivery methods 57.7 15 Individual procedures for each delivery method 34.6 9 Design-bid-build and design-build use the same method 7.7 2 Design-build and CM/GC use the method 0.0 0 Design-bid-build and CM/GC use the method 0.0 0 Total 26 TABLE 2 SURVEY RESPONSES REGARDING PROCEDURES FOR DIFFERENT PROJECT DELIVERY METHODS 2 5 5 9 9 10 17 0 2 4 6 8 10 12 14 16 18 No Improvement Needed Improved Usability Other - please specify Increased Automation Adaptability to Multiple Delivery Methods Improved Accuracy Increased Feedback/Communication Between Agency Divisions Number of Respondents Note: Respondents were asked to check all that apply FIGURE 20 Areas of improvement needed.