The Power of the Carrot: Using incentives to improve quality

Trenton M. Clark, P.E., President, Virginia Asphalt Association Kevin K. McGhee, P.E., Associate Director of Research–Pavements, Virginia Transportation Research Council

Psychologists have studied ways to motivate people for years. What methods work and do not work?

The value of using reward versus punishment to change behaviors is one of those methods. These approaches are continually being used in homes, schools, workplaces, etc. to have people conform, improve, or prevent actions.

For construction projects, the traditional approach was to conform (i.e., meet the specifications and requirements) or be punished (i.e., monetary reductions). Rewarding work above and beyond the minimum requirements was foreign to project owners. The common phrase was, “Why should I pay more to get what I asked for?” While valid on the surface, the owner’s mindset resulted in contractors striving no further than to meet the minimum requirements or specifications to get full payment. Seldom did contractors go beyond the minimum to potentially provide a better product or project. Absent alternatives to pay reductions or rejection, the quality outcome in a competitive context is a “race to the bottom.”

Initial Widespread Use of Incentives
Since the 1990s, the Virginia Department of Transportation (VDOT) has employed incentives based on a project’s final ride quality. Using a California-style Profilograph, VDOT technicians would push the device across a new pavement surface and assess the resulting ride against a specification. Meet the ride target, and the contractor gets full payment for that item; do better than what was expected, and receive a small bonus; but do worse, and expect a pay reduction. While a step in the right direction, the profilograph had many downsides such as practical applicability to more than new construction projects, maintenance requirements of the equipment, and the often poor correlation between the profile index and actual seat-of-the-pants ride quality. 

Realizing the challenges with the profilograph, the Virginia Transportation Research Council (VTRC) initiated a project in the mid-1990s, looking at the Inertial Road Profiler (Profiler) to replace the profilograph. The Profiler could be operated at highway speeds, construction and maintenance projects, and the output—IRI—better related to the roadway user’s ride. In 1998, VDOT began employing the Profiler- and IRI-based ride specifications on projects. As with the profilograph, the specification included an incentive along with a disincentive. However, the available incentive for providing a smooth ride remained relatively small.

In the early 2000s, VDOT and VTRC embarked on an effort to revise the IRI ride specification. By analyzing data and input by the paving industries, the bonus and penalty thresholds were adjusted based on project type (construction vs. maintenance) and roadway classification (interstate vs. non-
interstate). VDOT’s districts were highly encouraged to use these specifications. While there was reluctance by some VDOT personnel due to the potential “costs” of paying bonuses, VDOT Senior Leadership knew the improved ride would have long-term rewards. As shown in Figure 1 from the 2007 VDOT Ride Spec Report, ride quality was improving on paving projects in just a few short years.

To validate or refute the hypothesis that improved ride results in longer life and actual savings to the taxpayer, VTRC conducted a project and issued a report titled “Impact of a Smoothness Incentive-Disincentive on Hot-Mix Asphalt Maintenance Resurfacing Costs” in 2006. This report looked at the costs paid by VDOT in terms of bonus payments to contractors and the resulting ride quality. The researchers studied sites that were paved with and without the ride specification included. The report suggests that pavement life could be extended up to seven years and result in more than $1.3 million per year in direct savings. When particulars such as reduced fuel and vehicle repair costs, along with VDOT administration costs, are considered, the savings were much higher.

Since this initial effort, which included both incentive and disincentive provisions, the specification has evolved to include an incentive-only option. This option becomes relevant when project characteristics fail to meet the expected minimum criteria for the full ride spec requirements. However, if the contractor chooses to adjust their paving operations to exceed expectations, then the incentive-only option provides a mechanism to reward that extra effort. The incentive-only language was the first “carrot-only” approach to improve quality beyond the minimum acceptance limit.

Figure 1: Improvement in Ride Quality

New Carrots Popping Up
In late 2011, then VDOT Chief Engineer Mal Kerley established the Asphalt Quality Task Force. This task force was charged with developing what the Chief Engineer described as “…A simple, workable contracting mechanism that would help encourage quality.” This simple charge by Kerley led to various ideas and approaches to encouraging quality. By 2015, VDOT Chief Engineer Garrett Moore laid out three measures to promote quality and work timeliness. The first was related to the installation of permanent pavement markings on maintenance projects. While standard specifications would allow up to 30-days from completion of paving to install the permanent markings, Mr. Moore felt that having the markings applied sooner would substantially increase safety. To encourage this, Mr. Moore approved bonus language that would reward contractors who installed final markings within 14 or 21 days, depending on paving material (i.e. asphalt concrete or surface treatment). At the time, the pavement marking industry’s capacity was stretched thin, and the potential bonus was primarily designed to encourage new companies to enter the business or have existing companies expand. From the outset, this provision was given a 3-year active timeframe and then sunset. In the end, a few new companies started working in Virginia, and existing pavement marking contractors were able to increase staff and equipment. Overall, the carrot provided the desired result.

Along with the pavement marking bonus, Chief Engineer Moore approved the inclusion of asphalt density bonuses and asphalt mix consistency bonuses on pilot projects for 2016. While VDOT, VTRC, and VAA worked together to improve the performance of asphalt mixes through changes in gradations, volumetrics, and laboratory design compaction, Moore understood the value of consistent mixes that were properly compacted in the field. For the first time, VDOT included a 5% bonus for proper mix compaction and consistency along with a 5% bonus for tight control of the liquid AC during production. For pilot projects with these specifications, contractors would receive a bonus when the average density as measured by cores exceeded the minimum requirements, and at least 80% of the sub lots in a lot had a density above the minimum requirement. The addition of density bonus payments changed how many contractors approached a project. Contractors reported that they were more proactive in requesting patching before paving, pointing out concerns with scabbing after milling, using materials transfer vehicles, and being more attentive to weather forecasts and scheduling basic paving activities. While past approaches may have been good enough to achieve 100% payment, the “carrot” resulted in a paradigm shift.

What would VDOT get from this shift? As many research reports have documented, lower in-place density can have a detrimental impact on performance. The hope was for longer-lasting, better performing pavements through increased and more consistent compaction. After success with the 2016 pilot projects and widespread use in 2017–2018, VDOT and VAA agreed to adjust the disincentives and expand the situations where the bonus would be applicable. The percentage of disincentives was increased for surface mixes (See Table 1), and the minimum traffic threshold for bonus routes was lowered from 5,000 vehicles per day to 2,000. Both parties felt the change in disincentives would emphasize quality, and the lower traffic level would increase the potential impacts to more of the network. Figure 2 shows the difference in average compaction since 2017. VDOT’s dense graded mixes clearly show higher in-place values.

4492Note 1 – 2016 VDOT Road and Bridge Specifications Note 2 – 2020 VDOT Road and Bridge Specifications
Figure 2: Change in In-Place Density Since 2007

At the same time density bonuses were implemented, so was specification language related to asphalt mix consistency. VDOT specifications utilize a statistical approach to evaluate and monitor mixes during production. The average values for each lot of material (4,000 tons) are compared to acceptable limits through production tolerances on gradation sieves and AC content. If the lot’s average is within these limits for gradation and AC content, no production pay adjustment is assessed. Outside of these limits, payment adjustments are assessed on a project. Once more than one lot of material is produced, the material’s variability is calculated using standard deviation. Again, the standard deviation is applied to gradation and AC content results, and when the standard deviations are low, no pay adjustment is made. As the standard deviation exceeds thresholds, pay reductions are made. In extreme cases, the contractor can be required to remove out of spec, highly variable mixes.

With Chief Engineer Moore’s desire for more consistent mixes, VAA proposed specification language related to the standard deviation for AC content and gradation. By reviewing the 2014 contractor production data on surface mixes, VAA proposed a maximum standard deviation of 0.15 for AC content and maximum standard deviation limits for specification sieves. Additionally, no lot of material could be outside of the production tolerances. If all criteria are met, then the contractor would receive a 5% bonus for that lot. After meeting with VDOT, it was decided to concentrate on AC standard deviation and production tolerances for the 2016 pilot projects. In addition to the bonus, the penalty for variable mixes was also increased.

Figure 3: AC Standard Deviation
Figure 4: 2014 MITS PLAID Data

After the 2016 pilot projects, data for the 2017 to mid-2018 mixes were evaluated by VDOT to assess the AC consistency bonus’ impact. Overall, the effect was positive but not as dramatic as the field density bonus. Additionally, it was noted that some mixes were receiving bonuses for consistency, even when the final AC content was below the design value. The 2019 specifications addressed this concern with language that stated that the average AC content could not be more than 0.10% below the design value to be bonus eligible, along with the existing standard deviation language. This was done to ensure that sufficient asphalt was going into the mixes to improve durability.

Recently, VDOT shared an analysis of AC content standard deviation data from 2016 to 2019 (see Figure 3). In general, the trend lines from non-pilot 2016 mixes to bonus eligible mixes for 2017 to 2019 are similar. All four years had between 60 and 70% of the mixes meeting a standard deviation of 0.15. This analysis raised the question, “What value exists in this bonus?” It also raised the question, “Is the threshold for bonus correct?” As mentioned earlier, the 0.15 was based on the average standard deviation per approved job mix formula for the 2014 production data of VDOT’s surface mixes. A separate analysis of this data showed that approximately 60% of mixes (not tonnage) met the 0.15 (see Figure 4).

Overall, more analysis is required to answer both questions. Is there value in producing consistent mixes—most definitely! Less variability in good designs improves material life and overall performance. When a more uniform material is delivered to the job site, the paver and roller operators spend less time adjusting to variability (if they manage to adjust to it at all). An available bonus is an implicit recognition that superior consistency requires more effort for the asphalt plant operation from stockpile management to scale calibration to asphalt pump settings. These efforts come with a cost. In a low-bid environment, bonuses allow contractors to incur more upfront costs, anticipating at least recouping those costs with incentives.

Is the threshold set at the right value to reward contractors producing a consistent mix—the jury is still out! More evaluation of current and past production data is needed. A cursory review of the 2017–2019 data would suggest a threshold of 0.13, which is approximately the 50th percentile. The goal is to encourage contractors to produce a more consistent mix, understanding the challenges of asphalt plant production. Further evaluation of what is achievable is needed.

VDOT Senior Leadership has been a long-time proponent of bonuses related to paving. They have recognized the power of the carrot. Starting with rideability in the 1990s, VDOT has strived to reward exceptional work. Most contractors have benefited through the ride spec incentives, and so has the traveling public with smoother, longer-
lasting pavements. With the recent inclusion of density and mix consistency bonus payments, the benefits will be magnified. While the long-term pavement performance data does not yet exist, work by other agencies and researchers has pointed out the long-term benefits and reduced costs to agencies and users.

2007 VDOT Ride Spec Report (Figure 1)
VDOT Materials Division Presentation to the Paving Leadership Group on August 12, 2020 (Figures 2 and 3)
James Gillespie and Kevin K. McGhee, VTRC Report 06-R28 “Impact of Smoothness Incentive-Disincentive on Hot Mix Asphalt Maintenance Resurfacing Costs.” 2006