Alternate Design/Alternate Bid Process for Pavement-Type Selection (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.

36 CHAPTER SIX ALTERNATE DESIGN/ALTERNATE BID CASE STUDIES INTRODUCTION “MoDOT let 187 ADAB projects worth $2.2 billion. An average of 5.7 bidders competed for ADAB projects as opposed to 4.5 on conventional paving projects. MoDOT saw the average unit prices for HMA and PCC paving drop 5.1% and 8.6% respectively.” Ahlvers (2010) “The Indiana DOT recorded results in its SEP-14 ADAB projects with 4.3 bidders on conventional and 6.4 bidders on ADAB with overall savings of 9.0% in 2010 and 5.7% in 2011.” Duncan and Holtz (2012) Like most approaches to designing pavements, ADAB use differs from agency to agency based on individual experiences, climatic conditions, and traffic volumes. The previous chapters have chronicled the aspects of ADAB from a general practice level. This chapter will detail six case studies. Each was selected because it detailed a specific aspect of ADAB’s state of the practice. Kohn (1997) posits that case study research can be used for the following reasons: • to explore new areas and issues where little theory is available or measurement is unclear; • to describe a process or the effects of an event or an intervention, especially when such event affects many different parties; and • to explain a complex phenomenon (Kohn 1997). The following sections will describe a variety of agency approaches to ADAB in both the United States and Canada. In most cases, the case study will describe the agency’s process for ADAB, provide a practical example of a specific project, and furnish the results of the case study program as well as a short analysis of the advantages and disadvantages identified along with effective practices used by the agency. An attempt was made to collect a cross-section of the various approaches found in the survey. Many agencies use an LCC factor in the alternate bid program. Table 7 shows the summary of the case studies conducted for the synthesis. TABLE 7 CASE STUDY PROGRAM SUMMARY Case Study Agency Projects Rationale for Inclusion Remarks Indiana DOT US-31 (2009) plus 25 others let from 2010 to 2012 Uses an LCC bid factor on both HMA and PCC alternates in DBB and DB INDOT expanded its program and allowed contractors to submit bids for both HMA and PCC on the same project. Louisiana DOTD US-171/Gillis and US-171/Ragley (2003) plus 5 others let from 2014 to 2016 Uses A+B+C bidding Includes construction time factor Michigan DOT I-69 and US-31 Uses EUAC for LCC factor plus lane rental rates I-69 is a DB-Finance project; US-31 is DBB. Missouri DOT I-64 plus 3 DBB ADAB used on all projects Nation’s largest ADAB program Texas DOT Programmatic case study Data-driven ADAB decision tool Rational guidance for ADAB application Ontario Ministry of Transportation Highways 401 and 410 Uses stochastic LCCA to develop the LCC adjustment factor Primary user of ADAB in Canada

37 INDIANA DEPARTMENT OF TRANSPORTATION INDOT chose the US-31 at Kokomo, Indiana, project for its first ADAB project and constructed it under the SEP-14 program. In the SEP-14 application, the following goals were cited as the motivation for experimenting with ADAB: 1. Attract more bidders and competition, 2. Obtain true cost savings over similar conventional bid projects, and 3. Provide a more competitive market, that is, lower bid costs on paving items using this procedure versus the standard procedure where the pavement type is predetermined. (INDOT 2009) US-31 at Kokomo, Indiana, ADAB Project • Scope of Work: 2-mi section of a four-lane new alignment highway • Design Alternatives: Two alternatives evaluated through the MEPDG design methodology: (1) 14-in. HMA pavement; (2) 10-in. PCC pavement • Life-Cycle Cost Input Values: – 50-year analysis period – 5-year average unit costs from INDOT database – Discount rate = 4% – NPW analysis with future activities included is shown in Table 8. TABLE 8 INDOT LCCA INPUT VALUES HMA PCC Year Rehabilitation Year Rehabilitation 3 Crack Sealaa 8 Reseal Joints 6 Crack Seal 16 Reseal Joints 9 Crack Seal 24 Reseal Joints 12 Crack Seal 30 Grindb & Functional Overlay 15 Crack Seal 33 Reseal Joints 18 Crack Seal 36 Reseal Joints 20 Mill & Functional Overlay 39 Reseal Joints 23 Crack Seal 42 Reseal Joints 26 Crack Seal 45 Reseal Joints 29 Crack Seal 48 Reseal Joints 32 Crack Seal 35 Mill & Resurface 38 Crack Seal 41 Crack Seal 44 Crack Seal 47 Crack Seal Source: INDOT (2011). a Original table from INDOT (2011) indicates “Joint Seal.” b Original document states “Mill” Formula to Select Low Bidder: HMA Bid = HMA Contract Bid Amount + NPW of Future HMA Rehab PCCP Bid = PCCP Contract Bid Amount + NPW of Future PCCP Rehab • Outcome: 11 bids were received: five HMA and six PCC. Three contractors bid both HMA and PCC. The summary of the bids with the NPW factor for each pavement type is shown in Table 9.

38 As Table 9 shows, the LCC factor influenced the final outcome as the lowest HMA bid was below the lowest PCC bid price. Nevertheless, both low bids were below the engineer’s estimated cost for the pavement types, indicating that the increased competition resulted in reducing the proposed costs of both pavement types. TABLE 9 US-31 BID ANALYSIS SUMMARY PCC Initial Bid NPW Cost for PCC Bid Analysis A $11,273,863.10 $870,480.00 $12,144,343.10 B $11,734,858.25 $870,480.00 $12,605,338.25 C $11,882,813.21 $870,480.00 $12,753,293.21 D $12,489,884.99 $870,480.00 $13,360,364.90 E $13,049,350.43 $870,480.00 $13,919,830.43 F $13,380,701.06 $870,480.00 $14,251,181.06 HMA Initial Bid NPW Cost for HMA Bid Analysis B $11,098,853.08 $1,403,938.00 $12,502,791.08 C $11,342,588.69 $1,403,938.00 $12,746,526.69 G $12,047,454.98 $1,403,938.00 $13,451,392.98 E $12,547,049.20 $1,403,938.00 $13,950,987.20 H $14,670,217.40 $1,403,938.00 $16,074,155.40 Source: Adapted from INDOT (2011). INDOT 2010 and 2011 ADAB Program INDOT was pleased with the result because it achieved all three goals listed in its SEP-14 application. It then requested per- mission to extend the program into the next year and advertised 11 more ADAB projects, of which three were DB projects. The results of the 2010 annual ADAB program are summarized in Table 10. TABLE 10 2010 INDOT ADAB SUMMARY OF RESULTS Project HMA Bids PCC Bids Total Bids Winner Low HMA Bid ($M)a Low PCC Bid ($M)a Adjusted Low HMA Bid ($M)b Adjusted Low PCC Bid ($M)b LCC Impacted Low Bid PR 69/1 5 5 10 PCC 10.56 10.47 11.62 11.16 No US-31/3 4 5 11 PCC 23.46 21.87 25.4 23.07 No SR-25 5 7 12 PCC 27.6 26.7 31.06 28.8 No US-31/4 2 7 9 PCC 34.7 32.82 43.81 40.72 No I-70 2 5 7 PCC 33.6 32.33 38.69 35.87 No US-24 4 6 10 PCC 16.02 16.45 17.5 17.37 Yes PR 69/2 1 5 6 PCC 76.27 70.61 79.7 72.79 No PR 69/3 4 6 10 PCC 22.15 22.29 24.51 23.77 Yes PR 69/4c 5 7 12 PCC 39.95 40.6 43.39 42.69 Yes PR 69/5c 3 5 8 PCC 61.5 58.53 64.29 60.25 No PR 69/6c 1 4 5 PCC 86.1 83.9 91.37 87.4 No Source: Adapted from INDOT (2011). a Bid price without LCC factor; b Low bid + LCC factor; c Design-build project. As Table 10 shows, three out of 11 times PCC became the winning bid because of the influence of the LCC factor on the low HMA bid. Adding the US-31 project to the 2010 program makes the 2-year total show that 33% of the time HMA was actually the lower first cost, but when the LCC factor was added to the award algorithm, the PCC alternate was the winning bid 100% of the time. A joint FHWA and INDOT presentation (Duncan and Holtz 2012) summarized the results of the 2010 and subsequent 2011 ADAB program as shown in Table 11. The total savings calculated for the 2011 program was broken down as follows:

39 • “INDOT saved $3,800,000 immediately at the bid openings; • INDOT saved approximately $10,000,000 over the 50 year service life; • INDOT saved the taxpayers approximately $28,600,000 by being an additional 5.7% under the engineer’s estimate” (Duncan and Holtz 2012). TABLE 11 2010/2011 SUMMARY OF ADAB VERSUS CONVENTIONAL PAVEMENT BIDDING Year Bidding Type No. of Projects Total Winning Amount ($M) Total Engineer’s Estimate ($M) Percentage Below Engineer’s Estimate (%) Difference ADAB— Conventional (%) 2010 Conventional 19 285.3 345.4 17.4 9.0 2010 ADAB 11 422.7 574.2 26.4 Total Savings from ADAB $51,000,000 2011 Conventional 28 346.3 409.1 15.4 5.7 2011 ADAB 14 396.7 502.6 21.1 Total Savings from ADAB $28,600,000 Source: Adapted from Duncan and Holtz (2012). Although it could be argued that the savings below the engineer’s estimate were realized through conservative estimating practices, the fact that the percentage below the engineer’s estimate was reduced nearly by 30% indicates that the additional competition in the 2010 program as a result of ADAB was reflected in the bid tabulations used to complete the engineer’s estimates for the 2011 program. INDOT Case Study Analysis The previous discussion leads to the following conclusions about ADAB in the INDOT context: • ADAB increased the number of bidders. The SEP-14 report on the 2010 program indicated that the average number of bids for conventional projects was 4.32, whereas for ADAB projects the same average was 6.38 bidders on DBB projects and 5.33 on DB projects. • The ADAB winning bid amounts were lower than the winning conventional bid amounts by 9.0% in 2010 and 5.7% in 2011. This resulted in a cost savings of approximately $79.6 million below the engineer’s estimate. • The INDOT formula that includes an LCC factor for both pavement alternatives when determining the winning bid changed the final decision in roughly 33% of the cases. • In 67% of the cases, PCC won over HMA when bidding head-to-head without the LCC factor. This shows the value of allowing the market conditions at the time of letting make the pavement-type selection decision. • Assigning an LCC adjustment factor, the difference in NPV in future M&R costs, to the highest future M&R costs alternative does not appear to bias the final outcome. LOUISIANA DEPARTMENT OF TRANSPORTATION AND DEVELOPMENT LaDOTD chose the US-171 ADAB project and constructed it outside the SEP-14 program. The following goals were cited as the motivation for experimenting with ADAB: • “allow industry to select pavement type through the bid process; • enhance fair competition among paving industries; and • promote a more cost-effective use of the highway construction funds” (Temple et al. 2004). LaDOTD also believed that the “new process will reduce the tendency for selection determination based only on lowest initial construction cost” as well as encourage agency pavement designers to consider life-cycle aspects that have “the poten- tial to influence design strategy and construction sequencing with the objective of more efficiently managing traffic through the work zone” (Temple et al. 2004).

40 US-171 ADAB Project • Scope of Work: 4.1-mi section of a four-lane highway overlay • Design Alternatives: HMA Overlay versus PCC Bonded Overlay • Life-Cycle Cost Input Values: – 30-year analysis period for overlays (40 years is typical for new construction) – Current unit prices from LaDOTD database – Discount rate = 4% – NPW analysis with future activities included is shown in Table 12. TABLE 12 LADOTD LCCA INPUT VALUES FOR OVERLAYS HMA PCC Year Rehabilitation Year Rehabilitation 0 New AC Overlay 0 New Bonded PCC Overlay 15 Cold Plane & Overlay 15 No Action 20 No Action 20 Clean/Seal Joints 3 Patches Per Mile Source: Adapted from Temple et al. (2004). • Formula to Select Low Bidder: A+B+C Bidding HMA Bid = HMA Contract Bid Amount + Value of Time + NPW of Future HMA Rehab PCC Bid = PCC Contract Bid Amount + Value of Time + NPW of Future PCC Rehab • Outcome: Five bids (two HMA and three PCC) were received. The summary of the bids with the NPW factor for each pavement type is shown in Table 13. TABLE 13 US-171 GILLIS TO RAGLEY BID ANALYSIS SUMMARY PCC Initial Bid A Value Days B Value NPW Cost for PCC C Value Bid Analysis A+B+C A $18,132,887 305 $305,000 $736,000 $19,173,887 B $18,380,053 360 $360,000 $736,000 $19,476,053 C $19,042,639 500 $500,000 $736,000 $20,278,639 HMA Initial Bid A Value Days B Value NPW Cost for HMA C Value Bid Analysis A+B+C D $17,680,916 400 $400,000 $2,182,400 $20,278,639 E $18,707,278 500 $500,000 $2,182,400 $21,389,678 Source: Adapted from Temple et al. (2004). It can be seen in this case that the LCC cost factor influenced the winning bid. If this had been a standard A+B bid, the HMA low bid would have been roughly $400,000 lower than the PCC bid. A second ADAB project on US-171 from Longville to Deridder was let 3 months later and the low A+B+C bidder was the HMA alternative despite the LCC factors for the two options being roughly the same as those shown in Table 13. The winner’s HMA alternate bid portion was roughly $600,000 lower than the lowest PCC bid. The PCC bidder proposed 300 days to the 400 contained in the HMA bid. The winning edge was provided by the HMA contractor bidding $2.0 million less for the nonpavement work on the job. LaDOTD Case Study Analysis LaDOTD emphasized the schedule by including a value for time in the B-value and showed that including that value as a means to minimize user costs during construction can be successfully employed. Additionally, the results of the initial ADAB projects led LaDOTD to set a policy that if the estimated LCC of the two alternatives are within 20% of each other, then the

41 project will be let as an ADAB project. According to Brown (2010), LaDOTD’s policy is to use ADAB on most larger new construction and reconstruction projects. Five ADAB projects let between 2014 and 2016 were checked to determine if the trends seen in the case study still applied. Between five and seven contractors bid on each project. Three out of five bids went to the HMA bidder, and of the two PCC bids, one was decided by the impact of the LCC factor. The winning bid was between 2% and 8% below the agency’s estimate. Hence, it appears that the program’s original goals of increasing the number of bidders, letting the market decide which option is the most economic at the time of letting, and reducing construction costs over the long term have been achieved. MICHIGAN DEPARTMENT OF TRANSPORTATION MDOT constructed a number of ADAB projects from 2001 to 2008 through the SEP-14 program. In 2009, it received permis- sion to add the I-69 project to its original SEP-14 authorization. It cited the following goals as the motivation for including ADAB in the DBF (design-build-finance) project: • Permit both the concrete and asphalt industries to compete for this project; • Increase competition; and • Receive more favorable bids for MDOT” (Youngs and Krom 2009). I-69 ADAB Design-Build-Finance Project • Scope of Work: Reconstruct a 6-mi section of rural freeway, including the reconstruction of one interchange, pavement rehabilitation of an existing rest area, and rehabilitation of five structures. • Design Alternatives: HMA = 9.75 in. on a 6-in. aggregate base and 17.75-in. sand subbase, PCC = 10.25 in. on an 8-in. aggregate base and 11.75-in. sand subbase • Life-Cycle Cost Input Values: – 26-year analysis period, – Current unit prices from MDOT database, – Discount rate of 2.8% from current OMB guidance, and – EUAC analysis with future maintenance costs included at $63,713.76/lane mile for HMA and $38,353.40/lane mile for PCC. • Formula to Select Low Bidder: A+C with lane rental at $412.53/hour/lane closure HMA Bid = HMA Contract Bid Amount + Lane Rental + Future HMA Maint. costs (A/P, 2.8%, 26) PCC Bid = PCC Contract Bid Amount + Lane Rental + Future PCC Maint. costs (A/P, 2.8%, 26) • Outcome: Three PCC bids and no HMA bids were received. MDOT speculated that the additional earthwork required for the HMA option made it cost prohibitive and as a result, HMA contractors chose not to bid. US-31 ADAB Project By way of comparison, a 2011 MDOT ADAB project to reconstruct US-31 attracted five bidders including four PCC bids and one HMA bid. The results are shown in Table 14. Once again, MDOT speculated that the earthwork costs on the HMA option probably drove it to be the high bid. MDOT Case Study Analysis Like LaDOTD, MDOT also chose to emphasize the impact to road users during construction by including lane rental rather than using A+B bidding. The lane rental helps to incentivize accelerated construction schedules. Past research detailed in chapter three has shown that pavement LCCA outcomes are extremely sensitive to both the discount rate and the period of analysis (Pittenger et al. 2012). One study (Gransberg and Scheepbouwer 2010) found that if the discount rate was less than 4.5% and the period of analysis was less than 30 years, a bias favoring the PCC alternatives would always lead to lower PCC LCCs than equivalent HMA designs. Because both of these concerns apply to the MDOT algorithm, these two factors may explain why the ADAB bidding resulted in the selection of PCC as the most economical option in this case.

42 TABLE 14 US-31 BID ANALYSIS SUMMARY PCC Bid Price EUAC Lane Rentalb Bid Analysis A+C - Lane Rentala Ramps Freeway A $11,792,980 $725,768 $111,640 $413,502 $11,993,606 B $12,659,041 $775,229 $88,137 $635,140 $12,710,993 C $13,745,871 $837,869 $111,640 $625,216 $13,846,884 D $13,796,513 $840,190 $104,902 $527,629 $14,004,172 HMA E $15,475,645 $961,332 $111,640 $330,802 $15,994,535 Source: Adapted from Mikesell (2012). a Lane rental costs are subtracted from the EUAC adjusted Bid Price to arrive at total; b Hours bid for lane rental were not available. MISSOURI DEPARTMENT OF TRANSPORTATION MoDOT has been experimenting with ADAB since 1996 and is currently one of the few states that mandates ADAB on all projects over a certain size [greater than 7,500 square yards of continuous area (Roark 2011)]. The following goals were cited as the motivation for experimenting with ADAB: • Alternate bidding provides the opportunity for both asphalt and concrete contractors to bid on the two lowest cost designs head-to-head; • It also brings more contractors to the bidding arena, which translates into more competition and ultimately lower cost to the taxpayer; • [It] gives both industries the best opportunity to compete; and • Our customers benefit from “Best Value” choices. (Roark 2011) Although the first five ADAB projects did achieve the goals for the program, MoDOT identified a number of issues that needed to be addressed before proceeding. The primary lesson learned was to involve industry in the development of the ADAB process. According to one MoDOT senior manager, the initial ADAB approach was unpopular with several major stakeholders, including the contractors, pavement industries, and FHWA, owing to its perceived lack of transparency (Roark 2011). The agency also recognized that the 1986 AASHTO Design Guide did not account for local conditions and consistent inputs, which made it difficult for Missouri contractors to compete the two pavement types. MoDOT determined that before it could utilize ADAB on future projects, it had to meaningfully engage the design and construction industry, which it did. The industry outreach process moved forward “not necessarily with a hope of building consensus but rather in having industry at the table so that its representatives had a clear understanding of the process and how it was developed, [resulting in an] open and transparent process to ensure no bias toward one industry or the other” (Roark 2011). MoDOT adopted the MEPDG design method in 2005 (Ahlvers 2010) and made several ADAB policy decisions based on the results of the industry outreach efforts as follows: • ADAB to be used on all projects with a 7,500-square-yard continuous area of pavement or 14,000 square yards of discontinuous pavement (Brown 2010). The 7,500-square-yard area comprises approximately “a mile of paving and anything less than 7,500 square yards, the LCCA becomes small enough that it doesn’t make much difference in the bid” (Roark 2011). • The use of a 45-year analysis period for LCCA was agreed to. • Calibration of defaults in the MEPDG for creating equivalent structural design was accepted. • Fixed intervals for rehabilitation treatments that went into the LCCA were determined. • Grading design was conducted for each option’s thickness and clearly displayed in construction documents. • Consistency in pavement designs between different contracts to enhance bidding clarity for all MoDOT ADAB projects was agreed to. I-64 ADAB Design-Build Project • Scope of Work: Rebuild 10 mi of interstate highway west of St. Louis. Widen for an additional lane for a portion of the project. It is important to note that the DB procurement followed fixed price–best proposal procedures. Thus, the cost

43 of the project was capped at $535 million and the DB team that proposed the most scope for that price won. Contractors performed the pavement design alternatives. The project obtained a waiver of 23 CFR# 636.507(b) to allow the owner to reveal another offeror’s technical solutions. The selection was based on apparent best value and the waiver allowed for MoDOT to negotiate technical solutions of the unsuccessful proposer that were beneficial to the apparent best value proposer into the final contract. This allowed for the possibility of having technical solutions from both proposers. The results of each bid were significantly different, thus no technical concepts were shared between proposals; however, the use of ADAB in a DB framework was a key aspect for achieving the MoDOT goals for the project and both proposers believed the process was advantageous. • Design Alternatives: Supplied by competing DB teams based on mandated MEPDG design. Submittal requirements were as follows: – Pavement design method, including all design inputs – Design life – Rehab cycles required for design life provided – Pavement typical sections – Pavement and base thickness – Distress predictions – Minimum friction number. • Life-Cycle Cost Input Values: – 45-year analysis period – Current unit prices from MoDOT database – Discount rate = current OMB rate – No adjustment factor – Maintenance, salvage value, and user costs were not included in the LCCA (see chapter three for detailed discussion of the rationale). • Outcome: The project received two proposals, both of which proposed a PCC pavement system. The winning proposal included an 11-in. jointed PCC on 12 in. of aggregate base. MoDOT ADAB Practices in DBB Projects Because MoDOT mandates ADAB on most projects involving a significant amount of paving work, it is important to under- stand how its approach works on DBB projects. Table 15 provides an example of typical LCCA future maintenance and reha- bilitation schedules. Table 16 summarizes three typical DBB ADAB projects and shows that in only one case did the LCC adjustment factor change the order of the actual bids. TABLE 15 MoDOT LCCA INPUT VALUES HMA PCC Year Rehabilitation Year Rehabilitation 0 New Pavement 0 New Pavement 20 Mill & Functional Overlay – Mainlines 25 Diamond grind whole surface & full-depth repairs to 1.5% of surface area33 Mill & Functional Overlay – Whole surface 45 Remaining asset life = 0 45 Remaining asset life = 0 Source: Adapted from Ahlvers (2010). Roark (2011) indicated that out of 187 ADAB projects through 2010, the LCC adjustment factor had only reversed the deci- sion four times. Thus, it can be concluded that in Missouri, the head-to-head competition of HMA and PCC can be success- fully implemented purely on the basis of lowest bid if MoDOT chose to do so. The same reference furnishes a comparison of how the amount of work was split between the two pavement types when the LCC factor was not applied. • PCC total—$854,428,378 • HMA total—$871,075,824

44 • Difference—$16,647,446 (1.9%) • When the LCC factor is applied, the difference is 5.6%. TABLE 16 MODOT DBB ADAB PROJECT COMPARISON Project US-63 Adair US-36 Macon US-63 Randolph Adjustment factor (AF) $1,541,000 $964,800 $1,469,200 Low HMA bid $22,220,790 $40,499,627 $25,262,509 Low HMA bid + AF $23,761,790 $41,464,427 $26,731,709 Low PCC bid $24,320,546 $35,322,473 $26,452,184 Winner HMA PCC PCC AF impact on win No No Yes Source: Adapted from Roark (2011) and Fuerst (2014). One of MoDOT’s stated goals for implementing ADAB was to allow both the asphalt and the concrete industries to compete on every MoDOT paving job. The split in dollar terms appears to validate that the goal was achieved in an equi- table manner. One of the other goals was to benefit from the decreased pricing found in more competitive environments. A 2008 report (Rozycki and McCullough 2008) found that 2-year average unit prices for a ton of asphalt were $50.97 and $45.78 on conventional and ADAB project, respectively. The same trend was observed for PCC with unit prices of $139.59 and $126.75. Those changes represent a reduction in costs of roughly 10% on each product. Figure 14 shows the compari- son of the number of bidders on ADAB projects versus all projects and confirms that the goal to increase the number of bidders was also achieved. FIGURE 14 MoDOT bidder summary (Roark 2011). MoDOT Case Study Analysis The MoDOT experience has furnished a rich amount of ADAB experiential data that can be used to develop effective prac- tices that can be transferred to other agencies across the nation and around the world. The primary practices identified in the research are as follows: 1. Simplify the LCCA process to focus on the big-ticket items: construction cost and rehabilitation. The FHWA pavement LCCA manual, while showing that routine maintenance costs are properly included in the analysis, goes on to state that they can often be ignored as they are too small compared with the first cost to ever influence the outcome (Walls

45 and Smith 1998; Tinni 2013). MoDOT also dropped residual (salvage) value and user costs because both are highly theoretical and difficult to accurately quantify (Rozycki and McCullough 2008; Ahlvers 2010). 2. During its industry outreach, MoDOT identified that measurement and payment using the standard specifications created an equity issue when attempting to compete alternative pavement types (Roark 2011). Paying for asphalt in tons and concrete by surface area allowed the HMA option to recover the cost of every unit placed, whereas the PCC contractor got paid the same amount regardless of how many cubic yards of concrete were actually required as a result of surface irregularities. Thus, the alternatives did not share the same risk profile. This was resolved by measuring and paying for both options by surface area. 3. MoDOT chose to publish the LCC adjustment factor in the ADAB project solicitation to promote transparency of the procurement process. The literature shows that not knowing how the bid pricing will be impacted during bidding is an industry complaint in other states (Temple et al. 2004; INDOT 2009; Youngs and Krom 2009). 4. MoDOT set an agency standard for projects on which ADAB will be required. 5. MoDOT invested in a robust industry outreach program to ensure its ADAB program would be both well understood and accepted by its industry partners. ONTARIO MINISTRY OF TRANSPORTATION The Ontario MTO has been using ADAB since 2001 (Smith and Fung 2006). Its motivations were much the same as the state DOTs discussed previously (Tighe 2001). MTO has integrated the use of a stochastic LCCA model into its program, which avoids many of the inaccuracies of a deterministic model by allowing pricing to be evaluated over its historic range of values. MTO financed a study at the University of Waterloo that laid out the parameters for the LCCA model and discovered that the long-standing assumption that inputs to the classic LCCA are normally distributed was not true (Tighe 2001). The study showed that the lognormal distribution was a much better fit and changed the computed LCC. The difference is detailed as follows: Ignoring the lognormal nature of these variables introduces bias into an LCCA and in short does not reflect the true overall cost. This bias is further emphasized by the example presented whereby the normal distribution inputs result in an overdesign of the structure and subsequent increase in LCCA of approximately $62,000/km. (Tighe 2001; italics added) Highways 401 and 410 ADAB Projects • Scope of Work: Both projects involved the reconstruction of about 6 kilometers of a six-lane divided highway • Design Alternatives: HMA 11.8 in. on a 12-in. aggregate base; PCC 10.25 in. on 15.75 in. of aggregate base • Life-Cycle Cost Input Values: – 45-year service life – Current unit prices from MTO database – Discount rate = 5.3% – NPW analysis with future costs included is shown in Table 17. • Formula to Select Low Bidder: “LCC Advantage” TABLE 17 MTO LCCA ACTIVITY TIMING HMA PCC Year Rehabilitation & Maintenance Year Rehabilitation & Maintenance 0 New pavement 0 New pavement 8 Seal and patch 12 Reseal joints, partial depth repair 13 Seal and patch 25 Reseal joints, full depth repair, texturize 18 Mill and resurface 40 Reseal joints, full depth repair, texturize 23 Seal and patch — — 28 Seal and patch — — 32 Mill and resurface — — 45 Full depth base repair and resurface 45 Full depth base repair and resurface Source: Adapted from Holt and Hein (2011).

46 Adjustment Factor: NPW Future Asphalt Rehab – NPW Future Concrete Rehab Low bidder: Low bidder = lower of (PCC Bid Price + NPW Future Concrete M&R) vs. (HMA Bid Price + NPW Future HMA M&R) Outcome: The Highway 401 project received four bids, three PCC and one HMA; the Highway 410 project received six bids, one PCC and five HMA. The details of both are shown in Table 18. TABLE 18 MTO ADAB PROJECT COMPARISON Bidder ID Pavement Type Initial Bid ($M) LCC Factor ($M) Adjusted Bid ($M) Highway 401—Stage 3 A PCC 52.40 3.54 55.94 B PCC 53.34 3.65 56.99 C PCC 57.76 4.13 61.88 D HMA 60.05 3.85 63.90 Highway 410—Brampton A PCC 45.99 4.21 50.21 B HMA 50.07 4.27 54.34 C HMA 50.21 4.29 54.50 D HMA 50.65 4.33 54.98 E HMA 51.46 4.44 55.90 F HMA 53.11 4.63 57.73 Source: Adapted from Smith and Fung (2006). MTO Case Study Analysis MTO’s ADAB model is very similar to the one used by INDOT in that it adds the NPV of future M&R activities to both alternative pavement types. It should be noted that the LCCA protocol is published in the solicitation documents and then the actual construction cost is used to literally compute the adjustment factor for each bid during the tendering process. Thus, the initial bid price is used twice. The second time is to compute the salvage value. At this writing, MTO is reconsidering the value added by including the salvage value because it does not have a simple, noncontroversial way to validate the calculation (Holt et al. 2011; C. Raymond, “MTO ADAB Program Details,” personal communication, May 22, 2016). TEXAS DEPARTMENT OF TRANSPORTATION Texas DOT’s (TxDOT’s) ADAB program is included as a case study to inform the reader on a formal decision tool that assists agency personnel in determining whether to include ADAB in a specific project and, if ADAB is selected, to generate the pavement alternatives. This case study is not demonstrated for a specific project as no data were found that could be tied to the use of the tool. However, this case serves as a suitable illustration of a potentially effective practice that could be institutional- ized through a similar piece of software. TxDOT cited the following objectives as its motivation for implementing ADAB: • Bid[ding] pavements “head-to-head” allows open market to determine what is constructed; • Stop using assumptions made during the [pavement] evaluation/selection process years before letting; • Stretch our constrained funding as far as possible; and • Experience has shown alternate bidding increases the number of bidders [and] more competition keeps prices lower. (Lenz 2010) To support those goals, TxDOT sponsored a study conducted by the Texas Transportation Institute (Wimsatt et al. 2009) that developed a decision tool for identifying and selecting projects that would be likely to achieve the previous objectives if delivered using ADAB. The decision tool is called the Alternate Pavement Design Analysis Tool (APDAT) and it provides a methodology to screen pavement projects for the potential to accrue benefits through ADAB. Figure 15 illustrates the six-step logic used in APDAT.

47 FIGURE 15 TxDOT APDAT flowchart (Wimsatt et al. 2009).

48 The first step involves collecting project specific information and then filtering it through a series of Go/No Go parameters to determine whether to proceed. The initial filters are defined as project characteristics that make the project “unsuitable for offering alternative pavement design” (Wimsatt et al. 2009). Those factors are as follows: • This is a pavement widening project; • The project does not involve new construction or reconstruction; • The pavement is less than 500 feet in length; • The pavement is less than 5 miles in length, and both connecting pavements are either rigid or flexible pavements; and • There are areas of the pavement where truck traffic will be stationary for long periods of time. (Wimsatt et al. 2009) The tool includes a number of current and updated policies for TxDOT pavement design alternative development. The fol- lowing is a list of the technical features that are built into APDAT: • Use 30 years as the design life; • Use 4.5 and 2.5 for the initial and terminal serviceability values; • Use 95 percent reliability; • Use a time to first overlay of 15 years for the flexible pavement design; and • Design the overall subgrade and pavement structure to have a potential vertical rise no greater than 1.0 inch as calculated by Tex-124-E from soil tests in a soil column 15 feet deep as measured from the proposed finished pavement grade. Alternatively, provide material with an effective plasticity index of less than 25, to a depth of 8 feet from finished pavement surface. (Wimsatt et al. 2009) APDAT has the following LCCA input values: • Interest rate: 4% or other more appropriate rate. • Initial construction cost: Use of current TxDOT construction cost database. • Routine maintenance cost: Use data from the TxDOT Pavement Management Information System. • Future rehabilitations: Use data from the TxDOT Successful Pavements Database website. • PCC patching costs: Use data from the TxDOT Rigid Pavement Database. • User costs: Based on a sensitivity analysis, the following project-specific information is used: – Work zone time—day or night – Number of lanes open during construction – Maximum queue length – Work zone speed limit. • Salvage value: This input is optional. TxDOT Case Study Analysis The key point from the TxDOT case study relates to the goal of “using assumptions made during the [pavement] evaluation/ selection process years before letting” (Lenz 2010). As the input for APDAT illustrates, TxDOT is replacing assumptions with actual values that are derived from its pavement performance databases. The TxDOT LCCA model is the opposite of the MoDOT model, as it seeks to leverage the value of several decades’ worth of local data rather than simplify the LCCA. Therefore, the preceding case study analyses reveal that MoDOT and TxDOT represent distinctly different approaches in their LCCA methodologies. SUMMARY Conclusions One point that is not discussed is the impact of ADAB on the design costs of a paving project. MoDOT (Roark 2011), TxDOT (Lenz 2010), and MDOT (Youngs and Krom 2009) all included the incremental cost of designing two sets of pavements versus one in their calculation of actual savings. In all cases, the additional cost was found to be small compared with overall savings.

49 A number of conclusions can be drawn from the case study analysis and are described as follows: 1. In each case study, implementing ADAB increased competition by allowing both HMA and PCC paving contractors to bid on the same job. 2. The case studies documented more favorable pricing as a result of increased competition. MoDOT, the largest user of ADAB, recorded roughly a 5% to 9% decrease in the pricing for both pavement types. INDOT, LaDOTD, and MDOT all reported a decrease in pricing on their ADAB projects when compared with conventional paving projects. 3. In the majority of the cases presented, the LCC adjustment factor had no impact on the winning bid. MoDOT experi- ence stated the LCC factor affected the winning bid in only four of 187 lettings, whereas INDOT had the factor affect the award in three of 12 lettings. 4. ADAB has been successfully employed under a variety of project delivery methods, including DBB, DB, and DBF. Hence, this analysis has found that the applicability of ADAB practices should not be constrained by the type of proj- ects’ delivery methods. Effective Practices A number of effective practices were observed by means of the case study analysis and these are listed as follows: 1. A value of time component, such as A+B (LaDOTD) or lane rental (MDOT), can be added to the ADAB procurement to compete construction schedule between the two alternatives. 2. The LCC factor can be successfully applied to the HMA alternative only (LaDOTD, MDOT, MoDOT, TxDOT), to both HMA and PCC (INDOT), or not at all (MoDOT projects under 7,500 square yards). 3. Establishing a threshold, such as a percentage difference in LCC or a fixed amount of paving, removes the professional judgement issues about whether ADAB is appropriate for a given project. Said another way, the market value of each alternative is not known until the day of the letting. Thus, as in the TxDOT ADAB case, making an assumption about the market 1 or 2 years in advance risks missing a market movement that would result in better pricing if ADAB had been employed. 4. Simplifying the LCC factor input values reduces the amount of input data that must be collected to conduct LCCA. 5. Investing in industry outreach programs to ensure the ADAB program is both well understood and accepted adds value to the programmatic aspects of ADAB. 6. Maximizing both the clarity and transparency of the ADAB process by procedural actions such as publishing the value of the LCC factor was found to promote industry acceptance. Future Research As computing and data analysis power continues to increase, the use of DOT-specific, data-driven LCCA models similar to TxDOT’s is expected to overcome the shortcomings of the deterministic LCCA procedures that may be overly sensitive to the analyst’s assumptions. Leveraging that data to populate a stochastic LCCA model such as the model used by MTO further decreases the dependency on professional judgement to conduct LCCA. A future research project to assess the costs and ben- efits of developing a data-driven, stochastic pavement LCC model and comparing it with the simplified deterministic model in use by MoDOT would provide a comparison of the two approaches to LCCA. The objective of the proposed research should include identifying the most crucial input variables to satisfactorily estimate the LCC of pavement design alternatives. The outcome of the research would be a data-driven LCCA model and guidelines for a DOT to implement it if desired. The research would also make the business case for or against the investment in developing the requisite database that would be required to support the advance in technology. More important, the research would test the need to include LCC factors in ADAB at all in response to the MoDOT finding that the LCC only altered the winning bid four times in 187 lettings.