Public Benefits of Highway System Preservation and Maintenance (2004)

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.

8 CHAPTER TWO MEASURING PUBLIC BENEFITS Although precise definitions of the term maintenance and its component types vary among researchers and practitio- ners, a substantial literature presents the basic logic, func- tional models, and empirical evidence underlying the premise that PM and preservation are both efficient man- agement strategies. Over the past three decades, computa- tional tools have been developed, and to some extent stan- dardized, to assist highway system managers in estimating explicitly the cost-effectiveness of maintenance-based management strategies that entail PM activities. CHARACTERIZING MAINTENANCE ACTIONS The term “maintenance” refers broadly to any action in- tended to keep a facility or its parts functioning as origi- nally designed and constructed (Hudson et al. 1997) or, less restrictively, “the act of keeping fixed assets in accept- able condition” (Statement of Federal Accounting Stan- dards No. 6 . . . 1997)—that is, keeping conditions good enough, rather than at initial design levels. Agencies make expenditures when they carry out maintenance actions; these expenditures will typically be recorded as recurring costs in the LCCA. However, practitioners and even engineering textbooks do not agree on precise categories and definitions of activi- ties that constitute maintenance. One text suggests, for ex- ample, that maintenance actions may be classified as rou- tine, corrective, preventive, proactive, or reactive (Hudson et al. 1997). Another text distinguishes maintenance from “rehabilitation,” the former term restricted to actions that prevent or slow the onset of unacceptable service condi- tions, whereas the latter denotes action that returns unac- ceptable situations to acceptability, although the same text suggests that maintenance may be corrective or preventive, as well as routine or major, planned or reactive (Pavement Design . . . 1997). “Preservation” is not yet widely used by maintenance practitioners. As previously mentioned, Figure 1 (see chapter one) illustrates how some practitioners view the relationships of the terms. PM activities, a subset of maintenance, are planned and proactive (drawing on the AASHTO definition), in that PM “preserves the system, retards future deterioration, and maintains or improves the functional condition” (Pavement Preventive Maintenance Guidelines 2001). In building management, for example, managers may periodically re- place all light bulbs in an electrical display to lower the probability that the light bulbs will fail at critical times. Similarly, certain moving parts on commercial aircraft will be replaced after a particular number of flight hours, re- gardless of the part’s apparent condition. Within the context of LCCA, PM- and preservation- based management strategy will be judged appropriate— that is, efficient or cost-effective—primarily under two conditions: (1) for a new installation or substantial recon- struction, when it enables substantial reduction of initial costs (e.g., through materials and design details); or (2) for an existing installation in service, when the costs of re- sponding separately to problems as they seem impending or the losses when a problem does occur are high com- pared with the cost of the periodic maintenance action. In both cases, estimated net LCC reductions depend on the actual completion of planned maintenance programs and the validity of estimated characteristics of problems that may oc- cur if maintenance is not performed as programmed, as well as the frequency of those problems’ occurrence. PM, intended to prevent or delay problems, might seem to be intrinsically proactive, but practitioners use the term to refer also to reactive actions. For example, extensive sur- face cracking of a pavement may motivate managers to place a thin overlay rather than undertake crack sealing. Because the overlay both improves riding conditions and prevents water from penetrating beneath the surface and, as a consequence, probably defers for several years the need for crack sealing, the action of placing the overlay may be judged part of a PM-based management strategy. Crack sealing might, in this case, be considered normal or routine maintenance rather than as PM. However, as Table 2 illustrates, some highway mainte- nance analysts classify crack sealing as PM, presumably because the purpose of sealing is to forestall greater dete- rioration caused by water infiltration into the pavement’s base and subbase courses. Some practitioners assert that most pavement maintenance, even when described as preventive, is ultimately remedial (Research and Technology Coordinating Committee 1997), and clearly, crack sealing, microsurfacing, chip seals, and thin overlays are unlikely to be initiated unless there are superficial symptoms such as cracking to suggest that more serious problems could develop if action were not taken. Others (e.g., Zaniewski and Mamlouk 1999) suggest applying a chip seal to a pavement in good condition as a preventive action, and FHWA publications suggest that the distinction between PM and other maintenance is a matter

9 TABLE 2 CLASSIFICATION OF PREVENTIVE MAINTENANCE ACTIONS Types of Pavements Common Pavement Problem Preventive Maintenance Treatments Flexible Potholes Drainage Edge cracking Cracking sealing Lane-to-shoulder drop-off Slurry seal Aging Microsurfacing Thermal cracking Chip seals Thin hot-mix asphalt overlays Rigid Blow-ups Drainage Pumping Joint and crack sealing Joint faulting Retrofit load transfer Composite Potholes Drainage Edge cracking Cracking sealing Lane-to-shoulder drop-off Slurry seal Aging Microsurfacing Reflective cracking Chip seals Thermal cracking Thin hot-mix asphalt overlays Reseal sawed and sealed joints Source: Geoffroy 1996. of degree, referring to PM as “carefully timed, cost- effective treatments to roads experiencing only light to moderate distress” (Focus, June 2000, p. 1). The Texas Department of Transportation (DOT), as an example, defines PM projects as “work proposed to pre- serve, rather than improve, the structural integrity of the pavement and/or structure.” Examples include asphalt over- lays no more than 2 in. thick, seal coats, microsurfacing, cleaning and sealing of joints and cracks, patching of con- crete pavement, shoulder repair, scour countermeasures, cleaning and painting of steel bridge members, restoring drainage systems, cleaning and sealing of bridge joints, bridge deck protection, and more (Guide to Design Crite- ria 1999). The distinction between “functional condition” and “structural integrity” of the pavement or other highway component is a theme that is common in discussions of PM. Although explicit definitions of these terms and their distinctions vary among practitioners, they refer generally to the ability of the roadway to provide a safe and comfort- able trip versus the ability to withstand (i.e., without gross structural failure) the mechanical loads imposed by vehi- cles using the roadway. Currently used pavement and, to some extent, bridge design methods, however, typically use a pavement surface condition—that is, a functional con- dition—as a criterion or indicator of structural integrity, so it is unclear whether the distinction is truly useful. The Ohio DOT, for example, which defines structural integrity as the “ability of a pavement to carry anticipated loading,” is silent on the meaning of functional condition, but de- fines PM as “work performed on a structurally sound pavement . . . intended to preserve the pavement, retard fu- ture deterioration, and maintain or improve the functional condition without substantially increasing the structural capacity” (Pavement Design and Rehabilitation 1999). PAVEMENT CONDITION AND ITS INDICATORS If functional condition is not always explicitly defined, it al- most certainly has something to do with “service level.” This term, used widely in facilities management, refers to a meas- ure of how well the facility is able to perform the functions for which it was designed and built. By extension, service level often refers to the facility’s condition, on the principle that condition and function are highly correlated. The concept may be applied to facilities as a whole (e.g., an entire building, gas transmission system, or highway); however, typically it is applied to a single component (e.g., the roof, a pipeline, or a pavement), largely because of the difficulties of measurement when the intended functions have several dimensions. An early application of the service-level idea was the definition of pavement “serviceability” and development of a “present serviceability index” (PSI) to characterize the pavement’s ability to serve traffic. Ride quality was identi- fied as an important aspect of that ability, and the single- number PSI combined several measures of surface rough- ness and other observable physical characteristics of the pavement surface (The AASHO Road Test . . . 1962). That composite index was correlated with the judgments of rid- ers in vehicles traversing the pavement, as to whether the pavement’s condition was excellent, good, or poor, to yield what might now be termed the “standard model” for pave- ment performance (Figure 2). In this standard model, pavement condition or service level declines over time as a result of wear and damage in- duced by traffic loads and aging of materials. At some level of service (LOS), presumably defined by road users’ judgments that below this level the ride becomes poor, the condition is considered unacceptable. In the absence of any action to return the pavement’s serviceability to acceptable levels, the pavement has reached the end of its service life

10 FIGURE 2 The “standard model” of pavement performance. FIGURE 3 Example of the relationship between road-user costs and pavement condition (Source: Pavement Management Systems 1987). (although that service life is typically a period of time sim- ply chosen for analysis) (Haas et al. 1994). In principle, road-user costs increase as service level declines. With subsequent work, the PSI has been supplemented by a variety of other service-level measures, and ride qual- ity-based indices are generally used in highway manage- ment. No single measure has been generally accepted as a basis of pavement evaluation or design, although the Inter- national Roughness Index (IRI) may be the most widely used in the United States (Ksaibati et al. 1999). The various indices generally reflect engineers’ assump- tions that smoother pavements improve road-user riding comfort, vehicle safety, and operating costs. The indices themselves have typically been constructed to correlate with road users’ perceptions of ride comfort and stability (Liu and Herman 1998). Research has confirmed a strong correlation between rougher pavements and higher vehicle operating cost (see Figure 3). Surface roughness also has been shown to have a significant effect on single-vehicle and multivehicle crash rates (Karan et al. 1976; Al-Masaeid 1997).

11 Specific relationships between surface condition and road-user well-being, however, may not be straightforward. For example, a particular index may not apply equally well to all types of pavements (Wu 2000). Such parameters as gender of the observer and conditions under which the ob- servations are made significantly influence how individual observers judge the acceptability of a particular pavement’s condition (Chou and Wu 1997). Data collection methods and analysis procedures confound comparisons of service- level information from different agencies, even when the same index is used (Transportation Infrastructure . . . 1999). The variety of functions a highway is meant to provide further complicate the practical measurement of service level. Traffic operating conditions and congestion, for exam- ple, are measured by the aforementioned LOS rating (High- way . . . 1994). The LOS is correlated with travel speeds and vehicle delays, factors directly related to road-user costs, safety, and environmental quality. In contrast to pavement condition indices, the LOS grading system for traffic is used almost universally for design and management. Condition indices for highway bridges have been de- vised and are regularly reported, reflecting primarily bridge inspection observations that may indicate deteriorating structural integrity of the bridge’s superstructure, deck, and support structure. No single health index has been gener- ally adopted, however, although the National Bridge Inven- tory sufficiency ratings of load-bearing capacity must be periodically reported by states to the FHWA and are used in allocating federal funds for bridge maintenance activi- ties. The index employed in the Pontis bridge management computer program is popular; FHWA personnel report that 38 of 50 state DOTs use the Pontis program (Small et al. 1999). This synthesis review failed to find any widely used service-level indices either for highway drainage systems and appurtenances or for other types of facilities that might readily be adapted to highway system use. This failure is consistent with other reviewers’ findings (Hatry and Liner 1994; Transportation Infrastructure . . . 1999). MAINTENANCE AND THE SERVICE-LEVEL TRAJECTORY In the analysis of highway pavements and other compo- nents, service-level deterioration in the standard model is assumed to be a function of environmental conditions and loading, and its trajectory will be roughly as shown in Fig- ure 3. The model has gained wide acceptance as empiri- cally reasonable, but there is no theoretical or statistical ba- sis for specifying precisely a generally applicable shape of the service-level curve (e.g., Hudson et al. 1997; Prozzi and Madanat 2000). Within this model, maintenance is typically presumed to slow the rate of service-level deterioration or to increase service level or both. Some analysts presume that a mini- mum level of maintenance activity (e.g., normal or routine maintenance) is always implicit in the model, and failure to perform such tasks as, for example, cleaning drains, will increase the probability of damage and could accelerate observable deterioration rates. Unacceptable conditions would then occur sooner than otherwise expected (see Fig- ure 4). Geoffroy (1996) reports, for example, that studies by several agencies showed that such activities as crack sealing do reduce the amount of more serious deterioration observed later, and may add 1 to 2 years to the average time before the condition index reaches unacceptable lev- els. Some analysts imply that more substantial maintenance is required to avoid the precipitous onset of unacceptable service conditions, and suggest that certain key levels of the condition index should trigger various types of mainte- nance (e.g., see Figures 5 and 6). FIGURE 4 Effect of preventive or “normal” maintenance (Source: Lemer 1996). More typically, the standard model is used to illustrate the impact of actions (termed variously rehabilitation, re- pair, or reconstruction) that result in a substantial service- level increase and possibly a reduction of the deterioration rate. Such actions are presumed to rejuvenate the pavement (Figure 7). Some explanations seem to suggest that main- tenance, preservation, and rehabilitation are distinguished primarily by the size of the increase in condition index and the frequency of action (Figures 8–10). TYPES OF BENEFITS ATTRIBUTED TO MAINTENANCE Viewed within the context of the service-level standard model and LCCA, maintenance-based management strategies yield benefits by ensuring higher service levels and avoid- ing unacceptably low service levels, and doing so at a lower total cost than would be incurred if maintenance were not performed. These benefits, summarized in Table 3, accrue

12 FIGURE 5 Relationship of maintenance activities to stages of the service-level trajectory, example 1 (Source: Hicks et al. 1999). FIGURE 6 Relationship of maintenance activities to stages of the service-level trajectory, example 2 (Source: Pavement Management Systems 1987). to both highway agencies and road users, or occasionally to the public at large. Savings in the direct costs of providing good roads—that is, reductions in the net present value of total expenditures that an agency must make to keep its highways at acceptable ser- vice or condition levels—were the earliest and most immedi- ate benefits attributed to PM. Geoffroy (1996, p. 5), for exam- ple, cites a 1977 Utah DOT study finding that every dollar invested in PM early in the life of a pavement avoids expendi- tures of $3 later for major rehabilitation. Geoffroy cites other reports indicating that PM can extend the service life of port- land cement concrete pavements by 9 to 10 years and asphalt concrete pavements by 5 to 6 years, presumably reducing agency costs by deferring the need for more costly rehabilita- tion. Geoffroy cited some agencies that reported that PM re- duces the total time and money spent on pavement, compared with spending for on-demand maintenance activities only (i.e., making repairs when problems are reported) by 5% to 10% for pavements that have not yet been overlaid, and 16% to 20% for overlaid pavements. Such analyses, although possibly controversial (e.g., see Dasgupta 2001), have provided suffi- ciently convincing results that such states as Colorado, Michi- gan, and Pennsylvania have adopted explicitly identified PM as an integral part of their agencies’ highway programs (Galehouse 2002).

13 FIGURE 7 Effect of repair or rehabilitation, example 1 (Source: Geoffroy 1996). The Michigan Road Preservation Association, repre- senting the state’s contractors that specialize in such PM treatments as crack and joint sealing and surface seals for bituminous pavements and joint sealing on concrete pave- ments, for example, cites the FHWA in asserting that “for every dollar spent on PM to extend pavement life, a sav- ings of $6 to $10 can be realized.” The association also quotes a former director of the Michigan DOT: “This is a little bit of biting the bullet and spending the money on preventing problems, rather than 6 to 8 times more (money) to reconstruct or rehabilitate (the road) after the problem becomes serious” (“Introduction to Preventive Maintenance” 1999). Road users are presumed to realize savings through re- duced vehicle operating costs and reduced damage from crashes when roads are kept at higher LOS. Much of the FIGURE 9 Preventive maintenance as preservation strategy (Source: Optimizing Highway Performance: Pavement Preservation 2000). empirical support for these presumptions has come from studies of low-volume intercity roads (HDM Model De- scription . . . 1981; Pavement Management Systems 1987; Watanatada et al. 1987; Alfelor and Markow 1997; Pave- ment Design and Management Guide 1997). Widely used decision support tools (discussed in the next section) rely on these savings. Oglesby and Sargent (1962) concluded that reconstruction of older roads to new standards that were widely adopted in the 1950s and 1960s could not be rationally justified on the basis of direct fi- nancial and accident cost reductions alone. They recom- mended that inferred improvements in noncommercial ve- hicle operations should be included in the LCCA. Other benefits have subsequently been added to the analysis, in- cluding fuel savings, reductions in vehicle maintenance costs, pollution reduction, and other environmental en- hancements (Hallaq and Pettit 1982). Opportunities to realize savings by using roads offering higher LOS might be expected to influence road users’ travel behavior. Wachs (1967) reported that smoother FIGURE 8 Effect of repair or rehabilitation, example 2 (Source: Managing Public Infrastructure Assets to Minimize Cost and Maximize Performance 2001).

14 FIGURE 10 Oregon DOT public information presentation panel. TABLE 3 S UMMARY OF PUBLIC BENEFITS ATTRIBUTED TO PREVENTIVE MAINTENANCE OF HIGHWAYS Class of Benefit Recipient—Nature of Impact Problems and Limitations in Assessment or Realization Highway O&M cost reduction Agency—obviates or delays need for repairs, so that total expected agency expenditures over analysis period are reduced Savings calculation depends on forecast of need for repair. Savings calculation depends on assumed discount factor. Requires trade-off between present and future budgets. Service-life extension Agency—reduces deterioration rate or time- dependent probability of failure, so that anticipated need for repair is delayed Immediate impact of maintenance action may not be apparent. Service life is uncertain; failure may occur earlier or later than anticipated, regardless of maintenance. Service reliability improvement Agency—reduces statistical variability in observed service conditions, so that likelihood and frequency of unanticipated need for repair are reduced Requires substantial data to verify statistics and inferred correlation of service conditions with maintenance actions. Vehicle cost reduction System user—improves service conditions, so that road-user vehicle operating costs are reduced; workzone delays and hazards may increase costs Magnitude of savings depends on factors outside the control of transportation agency; e.g., vehicle fleet, traffic levels, assumed discount factor, assumed time values. Users may not fully perceive cost variations. No mechanism for agency to capture portion of savings; may actually reduce agency revenue. Ride quality/comfort improvement System user—improves service conditions so that user experience is enhanced; e.g., smoother ride, more pleasant views Recognition of improvement depends on factors outside the control of transportation agency; e.g., user demographics, traffic levels, vehicle characteristics. Traffic flow improvements System user—smoother surfaces enhance flow; workzone and detour delays and speed reductions and crash hazards occur during maintenance Directly perceived by road users. Safety improvement System user—improves service conditions, so that crash frequency or severity are reduced; workzone crash risk and expected severity may be greater Requires substantial data to verify statistics and inferred correlation of crash experience with service conditions and maintenance actions. No mechanism for agency to capture portion of savings. Environmental amelioration Public at large—directly or indirectly reduces stormwater runoff pollutants, air pollution emissions, noise; may be offset by pollution (e.g., sediment, herbicides) associated with PM Requires substantial data and sophisticated simulation models to verify inferred correlation of environmental conditions with service conditions and maintenance actions. No direct bases for estimating economic value of improvements. Note: Data are derived from the literature; see for example, Wachs 1967; NCHRP Synthesis of Highway Practice 58 . . . 1979; Fwa and Sinha 1986; Pavement Management Systems 1987; and Adams and Sianipar 1998. O&M = operations and maintenance.

15 pavement is among the characteristics of controlled-access routes that road users appreciate. Others have found that pavement roughness has an observable influence on vehi- cle speeds (Karan et al. 1976). Adverse impacts of maintenance activities on traffic conditions—such as workzone delays, hazards, and travel speed reductions—are definitely perceived as such by road users (Vadakpat et al. 2000). The review for the current synthesis found no instances of analyses that included these unfavorable impacts in LCCA comparisons of main- tenance-based management strategies. MANAGEMENT ANALYSIS OF MAINTENANCE BENEFITS Computer-based decision support tools have been devel- oped to help agencies devise management strategies that maximize the net benefits of a highway system. These tools have been based largely on LCCA principles—the service-level indices and the standard model described in the preceding sections. The World Bank’s Highway Design Model (HDM), for example, was an early and very popular example for pavement design and management (HDM Model Description . . . 1981). The HDM, widely used in- ternationally (Vincent et al. 1994), is now owned and main- tained by the World Road Association (PIARC). A similar model, MicroPAVER, was developed by the U.S. Army’s Construction Engineering Research Laboratory in coopera- tion with the American Public Works Association (APWA). According to APWA, MicroPAVER is used by cities and counties throughout North America. A number of consult- ing firms have developed similar but proprietary models. Those models explicitly represent the influence on road surface condition of neglect or maintenance performed, as a function of road design characteristics, traffic loads, en- vironmental conditions, and maintenance strategy (Pave- ment Management Systems 1987; Watanatada et al. 1987). They then project pavement conditions anticipated in the future, as a result of forecast vehicle loads and mainte- nance policies, and calculate estimates of road-user costs. Users of the HDM may attest to influence of different pavement designs and maintenance strategies on the net LCC. The HDM was intended initially for use primarily in de- veloping countries with lower-volume roads. A version of MicroPAVER has been developed for airfield pavement. Some agencies conduct testing or use their own collected observations to customize deterioration curves embedded in the available generic models (e.g., as reported by the TRB Data Analysis Working Group, an international forum for the discussion of methods of analysis of pavement per- formance data). In 1989, the FHWA issued a requirement—expanded by ISTEA (Section 1034)—that state DOTs must maintain some form of pavement management system to remain eli- gible for federal funds. FHWA officials report that the re- quirement was withdrawn in subsequent legislation, but many agencies continue to use such models in their pave- ment management activities. Bridge management tools based on similar reasoning have been developed as well, for example, under the aegis of AASHTO, through the AASHTOware program. The Pontis model is widely used, but the newer Bridgit model now offers an alternative tool. Using the approach embodied in these decision support tools, some state agencies—notably in Michigan and Cali- fornia (Pavement Preventive Maintenance Guidelines . . . 2001)—have attributed substantial financial benefits to their PM programs. The Michigan DOT reported having spent $80 million on the PM of 2,650 mi (4260 km) of highway since adopting its PM strategy in 1992, and it es- timates that $700 million of spending would otherwise have been required during the same period for rehabilita- tion and reconstruction projects to bring pavement condi- tions to comparable levels. More recently, the agency re- ported significant service life gains in both flexible and rigid pavement as a result of PM activities (Galehouse 2002). The Arizona DOT reported saving more than $200 million in maintenance and rehabilitation costs in the first 5 years after the agency implemented a pavement man- agement system (Madanat 1997). Federal officials reported that Michigan’s analyses have influenced other states to de- velop pavement preservation programs (“Forum II . . .” 2002). AGENCY USE OF PUBLIC BENEFITS ESTIMATES IN MANAGEMENT DECISION MAKING The survey of agencies conducted for this study confirms what the literature indicates: that use of pavement and bridge management software incorporating the concepts of public benefits of maintenance is nearly universal among state agencies (Table 4). A variety of customized packages are used for analysis of pavement maintenance strategies, but the Pontis model prevails in bridge management. Most agencies also use general maintenance management soft- ware (e.g., primarily work order processing) and fleet man- agement packages; the latter tools facilitate consideration of PM strategies in fleet operations. Widespread use of such analysis tools would seem to indicate that road-user costs are a significant factor in agency decision making. Many agencies (63%) have gone further by adopting construction bidding and award prac- tices—such as “lane rental” and “A+B” bidding—that

16 TABLE 4 SURVEY RESULTS OF HIGHWAY AGENCY USE OF ESTIMATED PUBLIC BENEFITS IN MAINTENANCE DECISION MAKING Uses of Public Benefits Percentage of Agencies Use pavement or bridge management software that embodies public benefits 94 Have used bidding or contracting methods that reflect road-user cost or other public benefits (e.g., “lane rental,” “A+B” bidding) 63 Use benefit–cost or LCC methods in maintenance management Use such methods, but only for major maintenance projects 37 21 Compare all maintenance with new construction in agency-wide programming and budgeting Make such comparison, but only for major maintenance projects 21 17 Report maintenance program accomplishments in terms of outcome measures (e.g., pavement condition) and report outcome information to the public 21 Use benchmarking in maintenance management and report benchmarking information to the public 21 make attention to road-user costs explicit to stakeholders outside the agency. Nearly three-quarters of agencies that had used such methods (approximately 47% of all agencies) claim also to have ongoing activities to identify public interests and concerns related to the agency’s highway programs. Such widespread adoption of practices that explicitly recognize road-user costs would seem to imply that agen- cies are particularly sensitive to the public benefits of vari- ous maintenance practices. Nevertheless, only one-third of agencies reported that life-cycle costing or other benefit– cost methods are used in their maintenance program plan- ning. Those that do use such methods in maintenance apply them for the most part only in assessing major projects (e.g., reconstruction). Only one-third of agencies reported that maintenance analyses are compared with new con- struction in agency programming and budgeting. Approxi- mately one-half of these agencies make the comparisons for major projects only. SUMMARY • The literature and surveys of current practice re- vealed three sets of issues pertaining to the measure- ment of public benefits of highway maintenance: • Definitions of maintenance, PM, and preservation— AASHTO publications offer definitions of the terms maintenance, preventive maintenance (referred to as PM in this report), and preservation; nevertheless, precise usage varies among practitioners and re- searchers. Determinations of what specific types of actions (e.g., overlays and crack sealing) qualify as maintenance rather than repair, renewal, or recon- struction vary as well. Maintenance activity overall is widely regarded as not adding significant capacity to the highway system or enhancing the structural ca- pacity of pavements and bridges; however, studies show that poor service levels (which maintenance could improve) may effectively reduce capacity. Highway management professionals are likely to classify as PM those actions that repair low levels of damage, to forestall the onset of more serious dete- rioration of service conditions. The lack of consistent definitions makes the comparison of various agen- cies’ PM practices difficult. • Maintenance impact on facility condition as a source of public benefits—Measures have been developed that characterize pavement and bridge conditions. Fa- cilities in better condition, as indicated by higher lev- els of these measures, are shown to be related to lower road-user costs, greater road-user comfort, safer traffic operations, reduced need for major main- tenance and repairs, and other benefits. Condition measures used to rate and manage facilities vary to some extent among agencies, although the IRI is widely used for pavements and the National Bridge Inventory sufficiency rating is periodically reported for larger bridges. An empirical model relating condi- tion to facility design characteristics, age and use, and maintenance strategy is widely accepted. That conceptual model has been used to devise computer- based decision support tools. Data used in these tools have been drawn from a few studies. • Estimating net benefits of maintenance—Arguments in favor of maintenance-based management strategies rely on principles of discounted cash-flow theory and LCCA procedures that embody those principles. LCCA is used to demonstrate that periodic agency spending for maintenance yields savings in agency construction, repair and rehabilitation expenditures, reductions in road-user costs, and other economic benefits, such that the total net LCC of providing adequate service conditions is lower than would be the case without such maintenance.