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.
Alternative Approaches to Heavy Vehicle Taxation Applications Manual - Evasion and Avoidance 8.0 Evasion and Avoidance There are two basic approaches~to estimating the overall evasion rate for highway taxes: inferring an overall evasion rate from evasion data obtained in the audit and enforcement process; and comparing actual tax collections to inclependent estimates of the amount that should be collected that are derived from data on highway usage. The first approach usually produces substantial underestimates of overall evasion; and, because of upward biases in currently used procedures for estimating truck VMI, the second approach usually produces substantial overestimates of evasion by heavy vehicles. The first section of this chapter discusses the first approach, the reasons why it usually underestimates overall evasion, and ways in which it can be modified to produce better estimates. Unfortunately, the modifications are relatively costly and may not be worth implementing. Accordingly, this section only presents suggestions for producing better evasion estimates and information that can be used to understand the extent to which relatively inexpensive applications of this approach are likely to underes~nate overall evasion. The second section of the chapter discusses the second approach to estimating evasion. That section presents brief descnptions of procedures for using reasonably unbiased VMT estimates to estimate evasion of highway taxes, and discusses limitations in the accuracy of the resulting evasion estimate. The revenue estimation procedures presume the use of improved procedures for estimating VMT by vehicle class, such as the procedures discussed In Section 3.~. The final section of this chapter presents a brief discussion of a related topic - estimating legal avoidance of highway taxes. 8.1 Use of Audit and Enforcement Information One method of estimating overall evasion of any tax or system of taxes is to infer an overall evasion rate from data on the extent to which evasion is detected through audit and enforcement activities. In the case of evasion that is detectable through audits, evasion rates for unaudited -firms usually are assumed to be the same as those for audited firms. In the case of other forms of evasion, such as bootlegging, the estimates of evasion are necessarily more judgmental. In general, evasion estimates derived from data on detected evasion are likely to uncleresUmate overall evasion because: Cambridge Systematics, Inc. 69
Alternative Approaches to Heavy Vehicle Taxation Applications Manual - Evasion and Avoidance I. Audits may not uncover ah evasion that is detectable through audits; 2. A lack of information on other types of evasion may result in underestimating their extent; and 3. Audited firms frequently have lower-than-average evasion rates. The following subsections discuss these limitations to the use of audit and enforcement information for estimating evasion along with ways of overcoming them. Unfortunately, overcoming these limitations can be relatively costly. Evasion that is Delectable Through Audits Several types of evasion are detectable through audits. These indude underreporting mileage in states with high taxes and/or fees and underreporting consumption of untaxed fuel. Failure to pay a tax on blending agents (such as kerosene) may also be at least partly detectable through audits. The overall extent of evasion that is detectable through audits can be inferred, at least in part, from the extent to which such evasion is detected in the course of auditing operators of vehicles subject to these taxes and fees. However, such inferences usually presume that ah such detectable evasion is actuary recognized in the course of these audits. To the extent that some of this evasion is not detected, the resulting estimates of the overall extent of these types of evasion will be low. Underreporting or misreporting of mileage by state poses an interesting set of detection Issues. It Is relatively easy to identify any inconsistencies between the mileage reports submitted for fuel-use reporting, registration-fee apportionment, and, if required, weight- distance tax payments. It is somewhat more difficult to identify inconsistencies between these reports and vehicle logs. And it may be even more difficult to identifier inconsistencies between the above information and other observations of a carrier's operations in a given state (obtained from vehicle observations at ports of entry, etc.~. Although auditors may be able to detect some under/misreporUng of vehicle mileage, detecting most or all such under/m~sreporUng requires a substantial effort. Furthermore, even when auditors working for a given state have both the necessary information and Incentives for detecting underreporting of mileage in their own state, they are likely to lack-both the- information- and incentives for detecting misallocation of mileage that adversely affects other states. For these reasons, it is likely that at least some evasion is potentially detectable through audits actually is undetected. Estimates of the extent of such evasion of-any--state'~-charg@s- require that a sample of-carriers be subjected to particularly thorough audits of operations in that state and that these audits extend to vehicles based In other states. In the case of fuel-tax reporting and registration fee apportionment, this last requirement poses an additional problem since both ihlA and Me IRP require that the audits be performed by auditors in the fleet's base state. 70 . Cambridge Systematics, Inc.
Alternative Approaches to Heady Vehicle Taxation Applications Manual - Evasion and Avoidance A particularly effective approach to using audits to estimate evasion was used in Me Oregon Weight-Mile Tax Study. About 35,000 observations of trucks, termed "snoops," were made by Oregon DOT field office staffs at random locations throughout the state over a period of several weeks. Precautions were taken to record license plate numbers and state without detection by drivers. However, since there was a significant amount of general awareness of the study, tax reports were carefully analyzed for the period covered to determine whether a higher level of tax payments were made. No enclence of such higher payments (i.e., recluced evasion) was detected from this analysis. The snoop ciata were then linked to the registered owners of the trucks through computer records, and sorted by firm. A sample of 217 firms, stratified by size of firm and groups of states, were then subjected to special audits. Normal audit procedures were used In the first part of each audit, uncovering as much evasion as is normally found during routine audits. After this, the snoop data were compared with Diver logs to find aciclitional travel not reported or uncovered during the Initial regular audit. The process used in the special audits performed in the Oregon study was expected to uncover almost all forms of evasion of the weight-mile tax except for night operations and operations over registered or declared weights. Additional field studies were used to estimate Me amounts of these other forms of evasion. Other Types of Evasion Types of evasion that are not readily detectable through audits inclucle bootlegging, operating a vehicle at (legal or illegal) weights that are higher than the maximum weight for which fees are being paid, or registering vehicles in incorrect categories in order to qualify for exemptions or tax-rate reductions. In states where diesel fuel is taxed at the rack and the tax rate is higher than in nearby states, bootlegging fuel into the state in truckload quantities may be a significant source of evasion.2 In order to estimate the extent of these forms of evasion, it is necessary to have an enforcement effort that is capable of providing some indication of the extent to which various evasion mechanisms are used and to extrapolate from this information to estimate overall evasion. One such source of information consists of stopping trucks in order to check their weight and their credentials. For purposes of estimating the extent of evasion, me only data used shouic! be data collected at sites3 and at Ones when drivers are not expecting to be Cambridge Systematics and Sydec with Pacific Rim Resources, Technical Memorandum No. 7 - Evasion Estimates Based on Audit Reports, prepared for the Oregon Legislative Revenue Office, the Oregon Public Utilities Comoussion, and the Oregon Department of Transportation, Salem, Oregon, February 1996. 2 Taxing diesel Fuel at We ~retaildevel~eliminates Me bootlegging problem, but it increases the number of taxpayers and the required enforcement effort. 3 There may be special interest in sites where evasion rates may be expected to be particularly high, such as weigh-station bypasses and urban-terminal access routes. However, if high violation rates are found at such sites, they cannot be assumed to apply to all trucks in Me state. Accordingly, although data from such sites are useful for identifying specific evasion problems, Hey may be of less value for He purpose of developing statewide estimates of evasion. Cambridge Systematics, Inc. 71
Alternative Approaches to Heady Vehicle Taxation Applications Manual - Evasion and Awidance stopped. We have found that, at temporary weigh stations, the percentage of trucks found to be violating tax laws ctrops appreciably after the station has been open for half an hour.4 Accordingly, for Me purpose of estimating evasion, only data collected from such sites during the first 15 or 30 minutes of operation should be used. Incorrect registration of vehicles includes: a) registering a truck used In for-hire operations as a private (non-commercial) vehicle; by registering a Muck or piece of construction equipment used-part of the dine on public highways as an off-~e-road vehicle; c) registering a bus used on a for-hire basis as a school bus; and d) registering a truck used for hauling agricultural products on a for-hire basis as a farm truck. This type of evasion often is clifficult to detect because the vehicles frequently are owned by firms or individuals that are not likely to be audited for highway user~harge purposes. If evidence of widespread violations of this type occurs, it may be desirable to expand the scope of firms and individuals coverer! by these audits or to expand the scope of audits performed for other purposes (e.g., income-tax audits) to Include checks for this type of evasion. Estimates of the extent of occurrence of this type of evasion can usually only be made juclgmentally, aided by asking questions of people In the field who may have firsthand knowledge of these types of operations. Extrapolation Issues In order to make efficient use of auditing resources, many states focus their motor-carrier field-audidng on large carriers based In state. Such carriers may expect to be audited on a regular basis. These carriers are likely to recognize that any readily detectable attempts to evade taxes owed to their base state may be futile. Accordingly, base-state auditors may detect relatively lithe evasion of taxes owed to their state by these carriers. The value of such auditing is likely to lie more in preventing evasion than In detecting it. The low rates of detectable evasion observed among large in-state carriers may not extend to smaller carriers or to those based in other states. Smaller carriers are likely to be audited less frequently, if at an; and they may fee! more confident in their ability to evade some of the taxes they owe. Also, they often are less aware of filing requirements and are frequently confused by the complexity of all the forms, regulations and laws. Similarly, out-of-state carriers may fee! they are able to underreport mileage in the state in question because, at least in the case of fuel taxes and registration fees, they win never be audited by persons from that state. If the state in question imposes relatively high taxes and fees, out-of-state carriers will have an incentive to underreport mileage in that state and overreport it in lower tax states. The above discussion indicates that both out-of-state carriers and small ~n-state carriers are likeIv to exhibit- higher' rates of detectable Evasion than larger ~n-state carriers. 4 Cambridge Systematics and Sydec win Pacific Rim Resources, Oregon Weight-Mile Tax Study, Final Report, prepared for me Oregon Legislative Revenue Office, Me Oregon Public Utilities Commission, and me Oregon Depallulent of Transportation, Salem, Oregon, February 1996, p.54. 72 Cambridge Systematics, Inc.
Alternative Approaches to Heavy Vehicle Taxation Applications Manual - Evasion and Avoidance Accordingly, estimates of the overall extent of detectable evasion that are derived by assuming that detectable evasion rates observed for audited carriers (which are mostly large instate carriers) are typical of rates for all carriers may substantially underestimate such evasion. Better evasion estimates can be developer! by stratifying the carriers and obtaining separate estimates of the average evasion rate for each stratum. One option would be to use four strata: small an~large in-state carriers; and small ancl'1arge 'out-of-state carriers. For each stratum, the average evasion rate should be estimated using data from audits of a reasonably representative sample of carriers In the stratum (and not just carriers that are relatively large or are known to be relatively easy to audit). Audits of carriers based' In other states (whether conducted by auditors employed by the state in question or by auditors employed by the carriers' base state) should include a substantial effort to identify any understatement of taxes owed to the state in question. If such an effort is not found to be practical, much evasion by out-of-state carriers may go undetected; accordingly, in this case, estimates of evasion by out-of-state carriers shouic! either be adjusted judgmentady or it should be recognized that the resulting arsenates of overall evasion will be low. S.2 Comparisons to Independent Estimates of Tax Liabilities A second approach to estimating overall evasion of any tax or fee is by developing an independent estimate of how much revenue the tax or fee should produce, comparing this estimate to actual revenue, and attributing any shortfall to evasion. Users of this approach to estimating evasion should be aware of two significant issues relating to the quality of the resulting evasion estimates: I. As discussed in Section3.l, the most commonly used procedures for estimating heavy-truck VMT generally overestimate this quantity, frequently by more than 25 percent. Use of the resulting overestimates of truck VMT will 1,c1,Allv no v~rv substantial overestimates of evasion. 2. J Rae ~ ~ J Even if better procedures are used for estimating truck VMT, modest errors in estimating the revenue that should be generated are unavoidable, and these errors are magnified when these revenue estimates are used to derive evasion estimates. Thus, if a given tax has a five percent evasion rate and the best available independent estimate of the revenue that should be generated by the tax is only accurate to ~5 percent, then The evasions rate estimated by this procedure may be anywhere between zero and ten percent. The first of these issues is more a near-term caution than an inherent limitation to the approach. As discussed in Section 3.l, procedures now exist for producing substantially better estimates of truck VMT with, at most, fairly minor biases. It is expected that, over time, most states will Implement these procedures and that further improvements will be made addressing the residual biases. Cambridge Systematics, Inc. 73
Alternative Approaches to Heady Vehicle Taxation Applications Manual - Evasion and Avoidance The second issue is an inherent limitation to the approach. However, in light of the difficulties that exist in using the previously clescribed approach to estimating evasion and the likelihood that overall evasion is underestimated when the earlier approach is used, this second approach may be of some interest. The remainder of this section presents procedures for using ~ndepenclent estimates of VMT to estimate the revenue that should be produced by apportioned registration fees, diesel-fuel taxes, and weight-c3istance taxes. The procedures presume the use of reasonably good estimates of statewide VMT by vehicle class such as those produced by procedures that are referenced in Section 3.~. In order to obtain completely independent estimates of the tax revenue that should be produced, the VMT estimates must be derived using procedures that make no use of information from the tax-collection process. (Some states currently adjust truck VMT estimates to be consistent with tax revenue.) An estimate of the total amount of revenue that shouic! be obtained from aU per-gallon fuel taxes or from just the per-gallon tax on diesel fuel and LPG can be obtained using the procedure descr~bec} in Section 3.3. Subtracting actual revenue obtained from these taxes produces an estimate of evasion of these taxes. Types of evasion that are reflected in this estimate include: Purchasing untaxed fuel and seDing it as taxed fuel; Selling fuel taxed in one state in another Higher tax) state (bootlegging); Blending untaxed fuel (e.g., kerosene) with taxed fuel and treating the blend as taxed fuel; Purchasing and using untaxed fuel without reporting it; Understating fuel use and overstating fuel economy in fuel-use reports; and · Underreporting mileage In high tax states (possibly combined with overreporting mileage in low tax stateside Also included- in this estimate will be some legal avoidance resulting from out-of-state fuel purchases by vehicles not subject to fuel-use reporting (though, in the case of diesel fuel, legal avoidance is fairly insignificant). In some states, the resulting estimate of evasion will be negative because of overreporting of mileage In the state (to take advantage of low tax rates in the state) and/or inaccuracies in the VMT estimates or other data used by the Section 3.3 procedure. Of somewhat less interest are estimates of the extent to which some states receive less than their share of apportioned registration fees as a result of misallocation of mileage among states. A state can estimate the extent of this form of evasion by: I) using the procedure presented in Section 3.2 to estimate the total amount of registration fees that should be paid by trucks with Free or more axles; and 2) subtracting the sum of actual 74 . . . . . . Cambridge Systematics, Inc.
Alternative Approaches to Heavy Vehicle Taxation Applications Manual - Evasion and Avoidance registration fee revenue from aU apportioned vehicles5 and from non-apportioneci vehicles with three or more axles. Although this evasion estimate may be of interest to some states, it reflects only one type of reg~stration-fee evasion,6 and nusaDocation of mileage also results in some states receiving more than their share of registration fees. Also, inaccuracies In the VIM estimates used by the Section3.2 revenue-estimation procedure could produce errors in the evasion estimates that are larger than the estimates. Finally, an estimate of evasion of a weight-distance tax can be obtained by using the procedure of Section 3.4 to estimate total revenue that should be obtained from this tax and subtracting actual revenue from the tax. This evasion estimate wig reflect evasion resulting from underreporting mileage operated in the state. However, it will not reflect evasion resulting from operating a vehicle at a weight that is higher than the maximum GVW declared for the vehicle or from misrepresenting vehicles as being entitled to partial or complete exemptions from We tax (for off-road use, etc.~. As observed at the beginning of this section, all of the above estimates of evasion presume He use of estimates of statewide VMT by vehicle class that are reasonably unbiased and Hat are not derived from revenue collection data; and, furthermore, even if such high- quality VMT estimates are used, small and unavoidable errors in the VMT arsenates may produce large errors In the evasion estimates. Hence, the resulting arsenates of evasion should be viewed as suggestive rather than definitive. · S.3 Avoidance Over the years, tax systems have been developed that minimize the opportunities for legal avoidance of highway taxes by operators of heavy trucks. Registration-fee apportionment and fuel-use reporting are two requirements that have been instituted in order to reduce opportunities for such avoidance. 5 The revenue from apportioned vehicles often includes a very small amount of revenue from apportioned two-axle sex-tire (FHWA Class 5) vehicles. If desired, a marginal improvement in this comparison can be made by subtracting an estimate of revenue received from apportioned Class 5 vehicles obtained by assuming that He revenue from these vehicles is approximately equal to He revenue Hat would be obtained from apportioned Class 5 vehicles based in the state if these vehicles operated only in He state. 6 Other possible types of registration-fee evasion are: Not registering a vehicle in He state as soon as required by law after it is brought into the state: ~ TO ~ _ 1 ~ · 1 . . . ~. · Using partial-year registration (which most states allow), but operating vehicles for a longer period of the year; · Registering at a lower GVW Han He maximum weight at which the vehicle is operated; · Registering as a publicly owned vehicle or over exempt type of use, and using the vehicle for private or commercial purposes; and Misrepresenting some tax-related characteristic of the vehicle, such as a resale price, or purpose for which He vehicle is used, such as private vs. for hire. Cambridge Systematics, Inc. 75
Alternative Approaches to Heavy Vehicle Taxation Applications Manual - Evasion and Avoidance One remaitung form of legal avoidance is the use of alternate routes to bypass toll stations on roads that use barrier tolls. The extent of this form of avoidance can be estimates] by comparing traffic counts (by vehicle class) on toll-road segments on which toll barriers are locateci with corresponding counts on free segments located before and after the toll stations. (The extent of such avoidance can also be reduced by introducing tolls on one or more exits prior to a barrier toll and on one or more entrances after such a toB.) Another form of legal avoidance is the minunization {or total avoidances of fuel purchases in jurisdictions with high fuel taxes. , ~ Fuel-use reporting requirements have largely eliminated this strategy as a legal means for heavy vehicles to reduce payments of state fuel taxes. However, the strategy can still be used by operators of vehicles that are not subject to fuel-use reporting (i.e., those with GVWs at or below 26,000 pounds), and it can be used by operators of all vehicles to avoid payment of sales taxes on fuel. Rough estimates of the avoidance of sales taxes on fuel can be estimated using the procedures presented previously for estimating evasion of per-gallon fuel taxes and the average price of fuel In the jurisdiction; however, in the case of local taxes, the VMT estimates used in developing these evasion estimates are likely to exhibit relatively large percentage errors when developed for small jurisdictions with relatively low VMT totals. As previously observed, for per-gallon fuel taxes, the evasion estimates produced using the procedures of Section 3.3 ant! S.2 will include (without differentiation) estimates of legal avoidance of these taxes. Also, since reg~stration-fee apportionment does not apply to trailers or to power units with declared GVWs at or below 26,000 pounds, the cost of registering those of these vehicles that are operated interstate can be legally minimized by registering them in states with low fees. One final type of legal avoidance is the modification of vehicles to qualify for lower fee rates or exemptions from fees. An example of this is boning camper bodies to pickups to avoid paying weight fees on pickups, as can be done legally in California, even if the campers can be easily removed for hauling other types of cargo. Similar modifications may be possible in other states to make otherwise "commercial" vehicles qualify as "non- commeraal" vehicles or similar categories having exempt or reduced-fee status. Estimates of the extent of occurrence of these types of avoidance can usually only be made judgmentalRy, aicled by asking questions of people in the field who may have firsthand knowledge of these types of operations. 76 Cambridge Systematics, Inc.
