MOTOR- VEHICLE GOODS AND SERVICES PRICED IN THE PRIVATE SECTOR
Report # 5 in the series: The Annualized Social Cost of Motor- Vehicle Use in the United States, based on 1990- 1991 Data
UCD- ITS- RR- 96- 3 ( 5)
Mark A. Delucchi
Institute of Transportation Studies
University of California
Davis, California 95616
madelucchi@ ucdavis. edu
www. its. ucdavis. edu/ faculty/ delucchi. htm
October 2004
ACKNOWLEDGMENTS
This report is one in a series that documents an analysis of the full social cost of motor- vehicle use in the United States. The series is entitled The Annualized Social Cost of Motor- Vehicle Use in the United States, based on 1990- 1991 Data. Support for the social- cost analysis was provided by Pew Charitable Trusts, the Federal Highway Administration ( through Battelle Columbus Laboratory), the University of California Transportation Center, the University of California Energy Research Group ( now the University of California Energy Institute), and the U. S. Congress Office of Technology Assessment.
Many people provided helpful comments and ideas. In particular, I thank David Greene, Gloria Helfand, Arthur Jacoby, Bob Johnston, Charles Komanoff, Alan Krupnick, Charles Lave, Douglass Lee, Steve Lockwood, Paul McCarthy, Peter Miller, Steve Plotkin, Jonathan Rubin, Ken Small, Brandt Stevens, Jim Sweeney, Todd Litman, and Quanlu Wang for reviewing or discussing parts of the series, although not necessarily this particular report. Of course, I alone am responsible for the contents of this report.
i
REPORTS IN THE UCD SOCIAL- COST SERIES
There are 21 reports in this series. Each report has the publication number UCD- ITS- RR- 96- 3 (#), where the # in parentheses is the report number.
Report 1: The Annualized Social Cost of Motor- Vehicle Use in the U. S., 1990- 1991: Summary of Theory, Methods, Data, and Results ( M. Delucchi)
Report 2: Some Conceptual and Methodological Issues in the Analysis of the Social Cost of Motor- Vehicle Use ( M. Delucchi)
Report 3: Review of Some of the Literature on the Social Cost of Motor- Vehicle Use ( J. Murphy and M. Delucchi)
Report 4: Personal Nonmonetary Costs of Motor- Vehicle Use ( M. Delucchi)
Report 5: Motor- Vehicle Goods and Services Priced in the Private Sector ( M. Delucchi)
Report 6: Motor- Vehicle Goods and Services Bundled in the Private Sector ( M. Delucchi, with J. Murphy)
Report 7: Motor- Vehicle Infrastructure and Services Provided by the Public Sector ( M. Delucchi, with J. Murphy)
Report 8: Monetary Externalities of Motor- Vehicle Use ( M. Delucchi)
Report 9: Summary of the Nonmonetary Externalities of Motor- Vehicle Use ( M. Delucchi)
Report 10: The Allocation of the Social Costs of Motor- Vehicle Use to Six Classes of Motor Vehicles ( M. Delucchi)
Report 11: The Cost of the Health Effects of Air Pollution from Motor Vehicles ( D. McCubbin and M. Delucchi)
Report 12: The Cost of Crop Losses Caused by Ozone Air Pollution from Motor Vehicles ( M. Delucchi, J. Murphy, J. Kim, and D. McCubbin)
Report 13: The Cost of Reduced Visibility Due to Particulate Air Pollution from Motor Vehicles ( M. Delucchi, J. Murphy, D. McCubbin, and J. Kim)
ii
Report 14: The External Damage Cost of Direct Noise from Motor Vehicles ( M. Delucchi and S. Hsu) ( with separate 100- page data Appendix)
Report 15: U. S. Military Expenditures to Protect the Use of Persian- Gulf Oil for Motor Vehicles ( M. Delucchi and J. Murphy)
Report 16: The Contribution of Motor Vehicles and Other Sources to Ambient Air Pollution ( M. Delucchi and D. McCubbin)
Report 17: Tax and Fee Payments by Motor- Vehicle Users for the Use of Highways, Fuels, and Vehicles ( M. Delucchi)
Report 18: Tax Expenditures Related to the Production and Consumption of Transportation Fuels ( M. Delucchi and J. Murphy)
Report 19: The Cost of Motor- Vehicle Accidents ( M. Delucchi)
Report 20: Some Comments on the Benefits of Motor- Vehicle Use ( M. Delucchi)
Report 21: References and Bibliography ( M. Delucchi)
Note that as of Spring 2004 reports 2 and 20 are not available.
There are several ways to get copies of the reports.
1) Most reports are available as pdf files on my faculty web page:
www. its. ucdavis. edu/ faculty/ delucchi. htm
2). You can order hard copies of the reports from ITS:
A. fax: ( 530) 752- 6572
B. e- mail: itspublications@ ucdavis. edu
C. ITS web site: http:// www. its. ucdavis. edu
D. mail: Institute of Transportation Studies, University of California, One Shields Avenue, Davis, California 95616 attn: publications
For general information about ITS, call ( 530) 752- 6548.
ITS charges for hard copies of the reports. The average cost is $ 10 per report. You can get a cost list before hand, of course. Or, you can have them send the reports with an invoice.
iii
3) The University of California Transportation Center ( UCTC) has posted Report # 1, the summary, on its website, as a PDF file. ( They might post more later). Go to “ Delucchi” in the alphabetical list at:
http:// socrates. berkeley. edu/~ uctc/ text/ papersuctc. html
4) FHWA, Planning Analysis Division, Office of Planning, 400 Seventh Street, S. W., Rm 3232, Washington, D. C., 20590, has a limited number of copies of Report # 1.
iv
LIST OF ACRONYMS AND ABBREVIATIONS AND OTHER NAMES
The following are used throughout all the reports of the series, although not necessarily in this particular report
AER = Annual Energy Review ( Energy Information Administration)
AHS = American Housing Survey ( Bureau of the Census and others)
ARB = Air Resources Board
BLS = Bureau of Labor Statistics ( U. S. Department of Labor)
BEA = Bureau of Economic Analysis ( U. S. Department of Commerce)
BTS = Bureau of Transportation Statistics ( U. S. Department of Transportation)
CARB = California Air Resources Board
CMB = chemical mass- balance [ model]
CO = carbon monoxide
dB = decibel
DOE = Department of Energy
DOT = Department of Transportation
EIA = Energy Information Administration ( U. S. Department of Energy)
EPA = United States Environmental Protection Agency
EMFAC = California’s emission- factor model
FHWA = Federal Highway Administration ( U. S. Department of Transportation)
FTA = Federal Transit Administration ( U. S. Department of Transportation)
GNP = Gross National Product
GSA = General Services Administration
HC = hydrocarbon
HDDT = heavy- duty diesel truck
HDDV = heavy- duty diesel vehicle
HDGT = heavy- duty gasoline truck
HDGV = heavy- duty gasoline vehicle
HDT = heavy- duty truck
HDV = heavy- duty vehicle
HU = housing unit
IEA = International Energy Agency
IMPC = Institutional and Municipal Parking Congress
LDDT = light- duty diesel truck
LDDV = light- duty diesel vehicle
LDGT = light- duty gasoline truck
LDGV = light- duty gasoline vehicle
LDT = light- duty truck
LDV = light- duty vehicle
MC = marginal cost
MOBILE5 = EPA’s mobile- source emission- factor model.
MSC = marginal social cost
v
MV = motor vehicle
NIPA = National Income Product Accounts
NOx = nitrogen oxides
NPTS = Nationwide Personal Transportation Survey
OECD = Organization for Economic Cooperation and Development
O3 = ozone
OTA = Office of Technology Assessment ( U. S. Congress; now defunct)
PART5 = EPA’s mobile- source particulate emission- factor model
PCE = Personal Consumption Expenditures ( in the National Income Product Accounts)
PM = particulate matter
PM10 = particulate matter of 10 micrometers or less aerodynamic diameter
PM2.5 = particulate matter of 2.5 micrometers or less aerodynamic diameter
PMT = person- miles of travel
RECS = Residential Energy Consumption Survey
SIC = standard industrial classification
SOx = sulfur oxides
TIA = Transportation in America
TSP = total suspended particulate matter
TIUS = Truck Inventory and Use Survey ( U. S. Bureau of the Census)
USDOE = U. S. Department of Energy
USDOL = U. S. Department of Labor
USDOT = U. S. Department of Transportation
VMT = vehicle- miles of travel
VOC = volatile organic compound
WTP = willingness- to- pay
vi
TABLE OF CONTENTS
REPORTS IN THE UCD SOCIAL- COST SERIES.................................................................... II
LIST OF ACRONYMS AND ABBREVIATIONS AND OTHER NAMES........................... V
TABLE OF CONTENTS........................................................................................................... VII
5.1 INTRODUCTION........................................................................................................... 1
5.1.1 Conceptual background............................................................................. 1
5.1.2 Cost items not usually included in GNP- type accounts of the cost of motor- vehicle transportation........................................... 4
5.1.3 Description of primary data sources......................................................... 4
5.2 THE ANNUALIZED REPLACEMENT COST OF THE MOTOR- VEHICLE CAR AND TRUCK FLEET............................................................................................... 9
5.2.1 The annualized replacement cost............................................................. 9
5.2.2 The cost of transactions involving used cars........................................ 10
5.2.3 Deduction for external replacement costs due to accidents................................................................................................... 11
5.3 THE COST OF MOTOR FUEL AND LUBRICATING OIL, EXCLUDING EXCISE AND SALES TAXES AND THE COST OF EXTRA FUEL USED BECAUSE OF TRAVEL DELAY..................................................................................... 12
5.3.1 Model of the cost of motor fuel............................................................... 12
5.3.2 Excess fuel consumption due to traffic delay ( parameter Ge)......................................................................................... 14
5.3.3 The pre- tax cost of gasoline and diesel fuel ( parameter Pa).............................................................................................................. 19
5.3.4 Producer surplus associated with motor fuels ( parameter PSF)....................................................................................... 20
5.3.5 The cost of automotive lubricants sold at retail.................................... 21
5.4 PARTS, SUPPLIES, MAINTENANCE, REPAIR, CLEANING, STORAGE, RENTING, TOWING, ETC., EXCEPT EXTERNAL COSTS OF ACCIDENTS.................... 21
5.4.1 The cost of automotive services.............................................................. 22
5.4.2 The cost of parts and supplies................................................................. 25
5.4.3 Deductions................................................................................................. 25
5.4.4 Estimating the annualized cost of long- lived repairs.......................... 26
5.4.5 Allocation to six classes of vehicles........................................................ 28
5.5 AUTOMOBILE INSURANCE: ADMINISTRATIVE AND MANAGEMENT COSTS, AND PROFIT................................................................................................... 28
vii
5.5.1 An estimate of the cost............................................................................. 28
5.5.2 Our estimate vs. the “ net premiums paid by persons” in the NIPA.............................................................................................. 29
5.5.3 Are automobile insurance prices optimal?............................................ 30
[ Note: there is no section 5.6]
5.7 PRICED PRIVATE COMMERCIAL AND RESIDENTIAL PARKING, EXCLUDING THE PARKING TAX................................................................................ 31
5.7.1 Priced private on- street parking............................................................. 31
5.7.2 Priced private off- street residential parking......................................... 31
5.7.3 Priced private off- street commercial parking....................................... 32
5.7.4 Total cost of private commercial and residential parking...................................................................................................... 33
5.8 TRAVEL TIME, EXCLUDING TRAVEL DELAY IMPOSED BY OTHERS, THAT DISPLACES PAID WORK................................................................................... 33
5.8.1 Background................................................................................................ 33
5.8.2 The cost per hour of travel time: concepts............................................ 34
5.8.3 Categories of travel, by type of vehicle, according to the data..................................................................................................... 34
5.8.4 Estimating the cost.................................................................................... 35
5.8.5 The cost of foregone monetary activity ( parameter Cm)............................................................................................................ 36
5.9 OVERHEAD COSTS OF BUSINESS, TRUCKING, AND GOVERNMENT FLEETS........................................................................................................................ 41
5.10 PRIVATE MONETARY COSTS OF MOTOR- VEHICLE ACCIDENTS............................. 41
5.10.1 Background.............................................................................................. 42
5.10.2 Methods used to estimate private monetary costs excluding user payments....................................................................... 42
5.10.3 Motor- vehicle user payments for the cost of motor- vehicle accidents inflicted on others..................................................... 44
5.10.4 Deducting automobile insurance administrative costs and property damage costs counted elsewhere as costs of motor- vehicle accidents........................................................... 45
5.11 DEDUCTION OF TAXES AND FEES INCLUDED IN THE PRICE- TIMES- QUANTITY ESTIMATES ABOVE.................................................................................. 47
5.11.1 Corporate income taxes.......................................................................... 50
5.11.2 Personal income taxes............................................................................ 50
5.11.3 Property taxes.......................................................................................... 50
viii
5.12 DEDUCTION FOR BUNDLED PARKING COSTS INCLUDED IN COST OF ANY INDUSTRIES ABOVE, BUT COUNTED SEPARATELY HERE AS A BUNDLED PARKING COST...................................................................................... 50
5.14 SUMMARY OF THE COST OF MOTOR- VEHICLE GOODS AND SERVICES PRICED BY THE PRIVATE SECTOR............................................................. 52
5.15 REFERENCES............................................................................................................. 53
TABLE 5- 1. DIRECT PAYMENTS FOR PERSONAL TRANSPORTATION, 1990- 1991 ( 109 $)......................................................................................................................... 63
TABLE 5- 2. DIRECT PAYMENTS FOR HIGHWAY FREIGHT TRANSPORTATION, 1990 AND 1991........................................................................................................... 68
TABLE 5- 3. EXPENDITURES ON MOTOR FREIGHT TRANSPORTATION IN SIC 421, 1991.................................................................................................................... 69
TABLE 5- 4. THE ANNUALIZED COST OF THE MOTOR- VEHICLE FLEET ( 1991 $)....................... 72
TABLE 5- 5. CALCULATION OF THE PRICE OF HEAVY TRUCKS, 1991......................................... 75
TABLE 5- 6. CALCULATION OF THE DEALER MARGIN ON SALES OF USED CARS, 1991........................................................................................................................... 77
TABLE 5- 7. THE COST OF MOTOR FUELS, 1991............................................................................ 78
TABLE 5- 8. OUR ESTIMATE OF TOTAL HIGHWAY DIESEL- FUEL CONSUMPTION IN 1987, COMPARED WITH THE FHWA’S................................................................ 80
TABLE 5- 9. GASOLINE AND DIESEL FUEL: COST, TAXES, AND RETAIL PRICE, 1987 AND 1991 ( CURRENT-$/ GALLON, EXCEPT AS NOTED).................................. 82
TABLE 5- 10. AUTOMOBILE INSURANCE PREMIUMS AND EXPENSES, 1991............................... 84
TABLE 5- 11. DATA ON WAGES AND TOTAL COMPENSATION, BY INDUSTRY AND OCCUPATION, IN THE U. S. IN 1990.................................................................. 86
TABLE 5- 12. ANNUAL COMPENSATION FROM TORT LIABILITY CLAIMS, CA. 1988 ( 109 $)................................................................................................................ 89
TABLE 5- 13. SUMMARY OF THE COST OF MOTOR- VEHICLE GOODS AND SERVICES PRICED IN THE PRIVATE SECTOR, 1991 ( BILLION $)................................ 90
FIGURE 5- 1. SUPPLY COST, PRODUCER SURPLUS, TAXES, AND FEES........................................ 92
FIGURE 5- 2. EFFICIENCY LOSS DUE TO MONOPOLY.................................................................. 93
APPENDIX 5- A: DATA ON WAGES AND COMPENSATION......................................... 94
ix
5. MOTOR- VEHICLE GOODS AND SERVICES PRICED IN THE PRIVATE SECTOR
5.1 INTRODUCTION
5.1.1 Conceptual background.
In this Report, I estimate the cost of motor vehicle goods and services priced in the private sector: the cost of the vehicles themselves, the cost of fuel and oil, cost of parts and maintenance, and so on. The economic cost of these motor- vehicle goods and services supplied in private markets is the area under the private supply curve: the value of the resources that a private market allocates to supplying vehicles, fuel, parts, insurance, and so on.
We do not observe the supply curve itself, and so cannot estimate the true private- sector resource cost -- the area under the supply curve -- directly. Rather, we must estimate this area indirectly, starting from what we can observe: total price- times- quantity revenues. Thus, the private- sector resource cost under the supply curve is equal to price- times- quantity revenues minus producer surplus and taxes and fees. We deduct producer surplus because it is defined as revenue in excess of economic cost, and hence is a non- cost wealth transfer from consumers to producers1. We deduct taxes and fees assessed on producers and consumers because in no case are they marginal- cost prices that can be used in a price- times- revenue calculation of costs2.
The relation between supply cost and producer surplus and taxes and fees is illustrated in Figure 5- 1. In that figure, the supply curve ( the private sector marginal- cost production curve), in the absence of fees and taxes, is S. A per- unit fee, such as the $/ barrel charge for the oil spill trust fund, shifts the supply curve up by a constant
1However, a net ( equilibrium) transfer from U. S. consumers to foreign producers is a real cost to the U. S.
2Recall that the point here is to estimate private- sector resource cost. The cost of the private- sector resources devoted to, say, making gasoline, does not include the federal and state gasoline tax, because that tax is a charge for the use of the roads, not part of the marginal- cost price of making gasoline. But, one might ask, why not then use the gasoline tax as an estimate of the cost of the roads, just as one uses price- times- quantity payments ( less producer surplus) to estimate private- sector resource cost? There are two reasons. First, we have data on expenditures on road construction and maintenance anyway, and so do not need to use price- times- quantity to approximate cost.
Second, even if we did want to use price- times- quantity to approximate the infrastructure cost, we would not use the gasoline tax for price, because it is not a marginal- cost price, but rather is a charge that bears no obvious resemblance to an efficient price. We can use price- times- quantity data to estimate cost ( the area under the supply curve) only if we know the relationship between price and cost. Because we do not know the relationship between the gasoline tax and cost, gasoline tax data are useless information in an analysis of cost.
1
$/ quantity amount, to Sf. A fixed- percentage tax, such as the sales tax, further shifts and also rotates the supply curve up, to Sft, such that at any Q the ratio of P at Sft to P at Sf is a constant.
Given the demand curve D and the final market supply curve Sft ( with the fees and taxes levied), Qft units are sold at price Pft to consumers and marginal cost Ps to producers. As mentioned above, we observe directly Pft and Qft, or their product Pft . Qft, but not Ps or the total cost as area under the private no- tax supply curve S ( this last area being what we wish to know). To get from the observed revenues Pft . Qft to the area under the supply curve ( 0- P0- a- Qft), we must subtract the revenues that are transferred to the government, and the revenues that are non- cost transfers from consumers to producers, as producer surplus. The government collects the difference between Pft and Ps as taxes and fees, in the amount Qft . ( Pft- Ps). Producers with costs lower than the marginal cost Ps collect producer surplus, equal in aggregate to the area P0- Ps- a.
Note that the result of this calculation is the cost actually incurred given prices and quantities as they were, not the cost that would have been incurred had there been no taxes in the first place. If there were no taxes and fees, the market price ( P*) to consumers would be lower, the marginal supply cost ( P8) would be higher, and the marketed quantity ( Q*) would be higher than in the actual case with taxes and fees. The resource cost in this case would be 0- P0- a*- Q*.
To worry, for example, about producer surplus is not merely a theoretical twiddle: it bears directly on comparisons of alternatives. For example, in comparing the cost of oil with the cost of alternative energy sources, it will not do to count all price- times- quantity revenues as the cost, because the true private resource cost is much less than this, on account of the enormous producer surplus that accrues to some oil producers.
The prices and quantities that obtain in private markets rarely are optimal -- that is, the actual prices ( P) paid rarely satisfy MSV = P = MSC -- not only because of distortionary taxes and fees, but because of imperfect competition, standards and regulations that affect production and consumption, price controls, subsidies, quotas, externalities, and poor information. For example, the market for crude oil is not always competitive. The reason, of course, is that the Organization of Petroleum Exporting Countries ( OPEC) sometimes manages to restrict oil output and thereby raise oil price above marginal cost. This is inefficient because it cuts off production of oil that could be produced for less than the [ formerly] prevailing market price and hence from a social- efficiency standpoint should be produced and consumed3 ( see Figure 5- 2). One also can
3This also results in an increased transfer of wealth from consumers to producers ( who are receiving a price above their marginal cost), and can be a real loss to heavy oil importers like the U. S. Note, though, that this extra wealth transfer is not in addition to price- times- quantity payments; to the contrary it already is part of price- times quantity payments. Rather, the extra wealth transfer is with respect to the total transfer in a competitive market ( see Greene and Leiby, 1993). The total resource cost of fuel use to 2
argue that other industries, such as the automobile manufacturing industry, at times look oligopolistic4.
Standards and regulations also can be economically inefficient. For example, the cost of vehicles and fuels includes items, such as catalytic converters and airbags and perhaps lightweight materials, used to meet government standards for emissions, safety, and fuel economy. Now, if the government standards are not the most efficient corrective, then the corresponding resources ( for catalytic converters, air bags, etc.) are not efficiently allocated. Of course, it is well known that, transaction costs and uncertainty aside ( and these admittedly are big asides), Pigovian taxes indeed are more efficient than are standards. However, Pigovian taxes can be more expensive to administer, less predictable, and more difficult to change on short notice, to point that standards might be preferable in some and perhaps many situations ( Baumol and Oates, 1988). It thus is not necessarily always the case that in the real world standards and regulations are less efficient than Pigovian regulations5.
Finally, consumers can be ignorant and irrational. For example, some and perhaps many people routinely underestimate the probability that they will be in an accident, and as a result undervalue safety equipment in motor vehicles.
In sum, we certainly do not have a dichotomous world of prices, in which private- sector prices are perfect and can be left alone, and all other prices ( or non- prices) need to be fixed. Rather, there are a variety of imperfections, in every sector, including the most competitive, unregulated private sectors, and hence a range of issues pertaining to pricing, taxation, regulation, and so on. We can be as concerned about the price of tires as the price of roads or the non- price of motor- vehicle emissions.
Price effects ignored. Note that my estimates of cost do not account for the affect on consumption of changes in price brought about by a hypothetical change in motor- vehicle use. For example, I in effect assume that if one reduces motor- vehicle use by 10%, the corresponding savings in motor- fuel will on average be 10%. However, the savings in motor- fuel actually will be less, because the price of motor- fuel will drop and
the U. S., competitive market or not, is equal to price- times- quantity payments less domestic producer surplus, which is a non- cost transfer from U. S. consumers to U. S. producers.
4In light of this, one might distinguish those resources provided in occasionally non- competitive markets, and place them in a separate column labeled “ subject to non- competitive pricing: msv = p ≠ msc”. For simplicity, I have not.
5I emphasize that the question here is not whether the resources required by government standards should be counted as a cost of motor- vehicle use -- they should be -- but whether they are efficiently allocated. Catalytic converters certainly are a cost of motor- vehicle use today, and barring unforeseen changes in regulations, will continue to be a cost of motor- vehicle use, regardless of whether or not there would be catalytic converters in a Pareto- optimal world. Furthermore, regardless of whether standards or taxes are used to address an externality, the relevant total cost is the resource cost of whatever control measures are used ( including “ defensive” behavior broadly construed) plus the estimated cost of the residual ( uncontrolled) effects, such as emissions.
3
thereby stimulate additional fuel consumption for the remaining 90% of motor- vehicle use. In principle, this problem arises no matter what the posited change in motor- vehicle use. For example, if one is estimating the total cost of all motor- vehicle use, one in principle should allow that the contraction in demand for steel would reduce its price and stimulate steel consumption in non- motor- vehicle sectors.
5.1.2 Cost items not usually included in GNP- type accounts of the cost of motor- vehicle transportation
Most of the cost items considered in this report show up in estimates by other analysts of the cost of owning and operating motor vehicles, or in the costs of motor- vehicle transportation in the National Income Product Accounts of the GNP. However, this analysis includes several items that most other analysts and most GNP- type accounts usually do not include. For example, the “ User Operated Transportation” categories of the National Income and Product Accounts ( NIPA) of the United States ( e. g., Bureau of Economic Analysis, 1990; Survey of Current Business, July, 1992), the FHWA’s Cost of Owning and Operating Automobiles, Vans, and Light Trucks ( 1984, 1992a), the U. S. Department of Labor’s Consumption Expenditure Surveys ( e. g., Bureau of Labor Statistics, Consumer Expenditures 1991, 1992), Runzheimers’ ( 1992) Survey & Analysis of Business Car Policies and Costs 1991- 1992; and the financial profile of automobiles in National Transportation Statistics ( 1992; their data are from the NIPA and the FHWA’s Highway Statistics) do not include in their accounts the following costs: compensated work travel time; the overhead expenses of business, commercial, and government fleets; accident costs paid for by responsible party, but not through automobile insurance; vehicle inspection by private companies; or the cost of legal services and security devices. They do not include them either because they have overlooked them, or because ( in the case of the NIPA and Consumer Expenditure Surveys) they classify them elsewhere, as legal costs, medical costs, housing costs, and so on, rather than as personal transportation costs.
There is no doubt, however, that these are costs of motor- vehicle use: for example there were no motor vehicles, there would be no vehicle inspection costs, and accident costs paid out of out pocket. The efficiency issue is whether or not motor- vehicle users recognize that these are costs of motor- vehicle use. That is, even though these costs are explicitly priced, they might be overlooked and omitted from the decision calculus. The out- of- pocket costs of motor- vehicle accidents might be an example of this sort of unaccounted- for cost.
5.1.3 Description of primary data sources
There are four primary aggregate estimates of ownership and operating costs of motor vehicles: 1) “ Personal Consumption Expenditures” ( PCEs) on “ User Operated Transportation,” in the National Income and Product Accounts ( NIPA) of the United States, estimated by the Bureau of Economic Analysis ( BEA) from basic data on economic activity in the U. S. ( Survey of Current Business, July 1992); 2) “ Consumer Expenditures” ( CEs) on transportation, estimated from a national survey of households,
4
administered by the Bureau of Labor Statistics ( BLS) ( BLS, Consumer Expenditure Survey: Integrated Survey Data, 1989; Division of Consumer Expenditure Surveys, 1993a); 3) The “ Nation’s Freight Bill, Highway” estimated by F. Smith ( Transportation in America, A Statistical Analysis of Transportation in the United States, , 1993), using data from the American Trucking Association, the Bureau of the Census, the Government Services Administration, and the Federal Highway Administration; and 4) The Census Bureau’s Motor Freight Transportation and Warehousing Survey: 1991 ( 1993), which surveys the revenues and operating expenses of firms that provide commercial motor- freight transportation6.
PCEs and CEs for 1990 and 1991 are presented in detail in Table 5- 1. Table 5- 1 also shows Smith’s ( 1993) estimates of personal and business expenditures for transportation. His estimates mainly are based on the PCEs but do include some original calculations. The PCEs are included in this table for comparison. Smith’s analysis of the “ Nations Freight Bill” is presented in Table 5- 2, and the results of the Census’ survey of trucking firms are presented in Table 5- 3.
The data of Table 5- 3 suggest that Smith ( 1993) might underestimate the nation’s freight bill. As shown in Table 5- 3, the total operating cost in SIC 421 ( trucking and courier services except air), excluding costs for purchased transportation, was $ 90 billion in 1991. ( We exclude all purchased transportation because purchased non- highway transportation is not relevant and purchased highway transportation would be double counted.) This however, covers only a fraction of commercial ( non- personal- use) trucking. If we assume that the ratio of the total operating cost to the fuel cost for all non- personal trucks is equal to this ratio in SIC 421, then we can scale the $ 90 billion in operating expenses in SIC 421 by the ratio of fuel purchased in SIC 421 to total fuel
6The General Services Administration ( GSA) of the U. S. Government also reports the operating costs of trucks ( in this case, trucks in large Federally owned fleets) ( GSA, Federal Motor Vehicle Fleet Report, for fiscal year 1990, 1993?), but because the total costs for each vehicle class are not broken down by type of cost, I cannot use the GSA data as the basis of any of my detailed cost estimates. However, the total operating costs can be compared with the equivalent total costs calculated from the Census data, in Table 5- 3.
In fiscal year 1991, Federally owned civilian light- truck fleets ( 8,500 lbs or less GVW) had a total operating cost of $ 0.25/ mile and Federally owned civilian heavy- truck fleets ( 24,000 lbs or more GVW) had a total operating cost of $ 1.03/ mile, and ( GSA, Federal Motor Vehicle Fleet Report, for fiscal year 1991, 1994?). The GSA operating cost includes depreciation cost, fuel cost, maintenance costs, and indirect costs of large Federally owned motor- vehicle fleets. The depreciation cost is estimated by GSA; the other costs are reported by fleet managers on Standard Form 82, “ Agency Report of Motor Vehicle Data ( Frisbee, 1994). On that form, maintenance costs include repair costs, preventative maintenance, motor oil, fluids, lubricants, replacement parts, and equipment ( such as cargo covers and fire extinguishers) needed to meet special operating requirements, and indirect costs include salaries of administrative and custodial staff, office supplies, building rental, utilities, tools and equipment, and capital improvements. Insurance is not included because the Federal government is self- insured, and registration fees are not included because the Federal government does not pay state registration fees.
In Table 5- 3 I calculate that the equivalent operating cost of trucks in SICs 4212 and 4213 is just over $ 1.00/ mile -- very close to the GSA’s figure of $ 1.00/ mile for heavy trucks.
5
consumed by all non- personal- use trucks. In 1991, firms in SIC 421 purchased 8 billion gallons ( Table 5- 3). On the basis of data in the 1987 Truck Inventory and Use Survey ( Bureau of the Census, 1990) and FHWA’s Highway Statistics 1992 ( 1993), we estimate that all non- personal- use trucks consumed 33 billion gallons of fuel in 19917. This suggests that all non- personal- use trucks had operating expenses on the order of $ 90 x 4.1 = $ 370 billion, or about $ 100 billion more than estimated in Table 5- 2.
( An alternative analysis, in which we separately scale local trucking ( SICs 4212 and 4214) and non- local trucking ( SICs 4213 and 4215), yields a similar result.)
BLS Consumer Expenditures. The Bureau of Labor Statistics ( BLS) surveys households across the U. S. to determine their expenditures on user- operated transportation. The surveys comprise an interview, in which householders report major purchases during the preceding three months, and a diary, in which householders record minor purchases ( BLS, 1988). The CE survey is administered to households only, not to any institutions, businesses, or government agencies. Expenditures include the full amount paid by consumers, including sales taxes and excise taxes. In the quarterly interviews the interviewer asks household members what percentage of transportation expenditures or vehicle mileage are for business use ( Bureau of Labor Statistics, Quarterly Interview Survey, 1991 Forms, Consumer Expenditure Survey, 1991), a question that suggests that transportation expenditures for business use are not counted.
BEA Personal Consumption Expenditures. The Bureau of Economic Analysis ( BEA) estimates Personal Consumption Expenditures ( PCEs) by type of expenditure, for “ User- Operated Transportation,” in the National Income and Product Accounts ( NIPA) of the United States ( Survey of Current Business, “ National Income and Product Accounts, 1992). According to the BEA, “ persons” consist of individuals, nonprofit institutions, private noninsured welfare funds, and private trust funds, and PCEs include goods and services purchased by individuals, the operating expenses of nonprofit institutions, and the value of food, fuel, clothing, rent, and financial services received in kind by individuals ( BEA, Personal Consumption Expenditures, 1990; Byrnes et al., 1979). PCEs exclude the following: expenditures by businesses and by the government, including reimbursable business expenses by persons and expenses related to the business use of motor vehicles purchased for both business use and personal use; traffic fines, parking fines, motor- vehicle registration fees and driver’s- license fees, which are included under “ Personal Tax and Nontax Payments” in the
7Our analysis of the 1987 Truck Inventory and Use Survey ( Bureau of the Census, 1990) indicates that personal- use trucks consumed 43% of total fuel consumed by all non- public trucks ( the Census TIUS data do not include public vehicles). In 1991, all trucks, including public trucks, consumed 56.8 billion gallons of fuel ( FHWA, Highway Statistics 1992, 1993). Comparing FHWA estimates of VMT in 1987 with the TIUS estimates ( the FHWA estimates include public trucks), we estimate that public trucks constituted 4% of all truck VMT ( FHWA, Highway Statistics 1992, 1993). Assuming that they also constituted 4% of total truck fuel consumption, then all non- public trucks consumed about 54.6 billion gallons in 1991. If 57% of this amount was for non- personal use, then private commercial ( non- personal- use) trucks consumed about 31 billion gallons of fuel. Adding the 2 billion gallons consumed by public trucks yields a grand total of 33 billion gallons of fuel consumed by all non- personal- use trucks.
6
NIPA; finance charges, which are counted as “ Interest paid by Consumers to Business”; and transactions between individuals, such as the sale of a car from one person to another ( such transactions cancel out); ( BEA, Personal Consumption Expenditures, 1990; Byrnes et al., 1979). Generally, the BEA makes detailed estimates of PCEs every five years when the Census publishes its quinquennial economic censuses of agriculture, transportation, manufactures, wholesale trade, retail trade, service industries, construction industries, mineral industries, and governments. These detailed estimates are called “ benchmarks”. In non- benchmark years the estimates are less complete, and are made partly by extrapolation, interpolation, and judgment. The BEA uses data from the Bureau of the Census, other government agencies, trade organizations, and other sources, as well as its own judgment, to estimate total expenditures on transportation and to allocate the total to business, government, and personal use. For details, see Byrnes et al. ( 1979) and especially the BEA ( Personal Consumption Expenditures, 1990).
The PCEs are meant to include all sales taxes paid, including local taxes on parking ( Key, 1993), but it is possible that in some cases, unbeknownst to BEA, the source data that the BEA uses do not include relevant taxes.
Discussion. These capsule descriptions indicate that the coverage of the BLS’ CEs differs from the coverage of the BEA’s PCEs in at least one way: the PCEs include expenditures by non- profit organizations, whereas the CEs do not. Differences in definition and estimation of individual expenditure items are discussed below. ( See also the Division of Consumer Expenditure Surveys, 1993b).
The PCEs and the CEs are estimates of personal or household expenditures on transportation, and Census and the TIA estimates are of costs or revenues of motor carriers. This distinction between personal, business, and commercial transportation is unfortunate for me, because in most cases it is irrelevant to the classification and analysis of the economic costs of motor- vehicle use. Whether or not a particular vehicle carries people instead of goods, or is “ personal” or for “ business” or “ commercial” purposes has nothing to do with the amount of pollution it generates, the amount of road damage it causes, the amount of public service that it “ consumes”, and so on. It also has little to do with the amount of taxes and fees it is assessed. On the other hand, the kind of fuel that a vehicle uses, the amount that it weighs, and whether or not it is a truck, have a lot do with the costs that it engenders and the taxes and fees that it pays ( Federal Highway Administration [ FHWA], Highway Taxes and Fees, 1991). Therefore, I eschew the personal/ commercial distinction, and instead distinguish between gasoline and diesel fuel, and between three size classes of classes of vehicles, ending up with six vehicle types8:
8Those interested in seeing a breakdown of travel by household vehicles, business- use vehicles, commercial light and heavy trucks, government vehicles, buses, and other vehicles, see Table 4- 1 of Report # 4. My analysis there, and in the table presented below, indicates that business- use VMT is about 30% of personal- use VMT. One can get a rough idea of the extent of business- use travel by comparing data on personal use of passenger cars in 1991 with total travel by passenger cars in 1991:
7
• Light- duty gasoline automobiles: passenger vehicles, including station wagons and motorcycles, that use gasoline as a fuel. In some cases I ignore motorcycles, which account for but a tiny fraction of highway travel ( FWHA, Highway Statistics 1992, 1993) and emissions ( EPA, National Air Pollutant Emission Trends, 1900- 1994, 1995).
• Light- duty gasoline trucks: trucks, vans, minivans, jeeps, and utility vehicles, that run on gasoline and have a gross vehicle weight rating of 8,500 lbs or less and a curb weight of 6,000 lbs or less. ( The FHWA’ annual Highway Statistics annual report uses a slightly different category, “ two- axle, single- unit” trucks.)
• Heavy- duty gasoline vehicles: all other trucks, and buses, that run on gasoline. In some cases I ignore buses, which in the U. S. account for a tiny fraction of highway travel ( FWHA, Highway Statistics 1992, 1993).
• Light- duty diesel automobiles: same as light- duty gasoline automobiles, except that they use diesel fuel.
Number of passenger cars in 1991
VMT by passenger cars in 1991
Gallons used by passenger cars in 1991
( millions)
( billion)
( billion)
Personal use ( RTECS, 1991)
108.3
1,150.0
54.5
Personal use ( NPTS 1991)
122.6
1,135.2
not reported
All uses ( FHWA 1991)
142.6
1,533.6
70.6
The triennial Residential Transportation Energy Consumption Survey ( RTECS) measures VMT and energy consumption by households for personal transportation ( EIA, Household Vehicles Energy Consumption 1991, 1993). It covers only vehicles that are kept at home and are available for “ some” personal use ( p. 164). It excludes motorcycles, mopeds, large trucks, and buses, but includes company- owned vehicles that are “ ordinarily” kept at home and “ regularly” available for personal use ( p. 221). It also includes household vehicles used for job- related activities. The latest data ( shown here) are for 1991.
The Nationwide Personal Transportation Survey ( NPTS), conducted every seven years, surveys personal travel in the United States. It includes cars, trucks, vans, RVs, motor homes, motorcycles, and mopeds “ owned, or available for regular use” by household members ( Federal Highway Administration, User’s Guide for the Public Use Tapes, 1990 Nationwide Personal Transportation Survey, 1991; survey form p. 4). Thus, the coverage of the NPTS probably is very similar to the coverage of the RTECs. The latest data ( shown here) are from 1990; I have extrapolated to 1991, using 1991/ 1990 ratios from FHWA VMT data.
The Federal Highway Administration ( FHWA, Highway Statistics, annual) reports total VMT by all passenger vehicles, regardless of use, on the basis of traffic counts on roads. The data shown are for 1991.
Again, these data, and the similar analysis of Table 4- 1, indicate that business use of motor vehicles ( which is not the same as commercial use) is at least 30% of personal use.
8
• Light- duty diesel trucks: same as light- duty gasoline trucks, except that they use diesel fuel.
• Heavy- duty diesel vehicles: same as heavy- duty gasoline vehicles, except that they use diesel fuel.
Because the published primary- source estimates of ownership and operating costs ( mentioned above) pertain to personal- use or commercial use vehicles, rather than to my six classes of vehicles, I must adapt the published estimates to my six vehicle classes, or dig deeper into the underlying source data. In the following sections I detail my estimates of the cost of vehicles, finance charges, fuel and lubricants, maintenance and repairs, parts, and insurance, in each of the six classes described above.
5.2 THE ANNUALIZED REPLACEMENT COST OF THE MOTOR- VEHICLE CAR AND TRUCK FLEET
5.2.1 The annualized replacement cost
We estimate the annualized replacement cost of the motor- vehicle fleet as:
ARC= i×I1− 1+ i( )− tI= Q×C eq. [ 5- 1]
where:
ARC = the annualized replacement cost ($/ year)
I = the total investment, or complete replacement cost ($)
i = the annual interest rate for investment in motor vehicles
t = the term of the investment in motor vehicles ( years)
Q = the present quantity of motor vehicles ( unregistered as well as registered)
C = the present cost per motor vehicle ( excluding producers surplus, taxes, and fees)
Note that equation 5- 1 annualizes the entire replacement value at t= 0, which means conceptually that the entire vehicle stock is replaced overnight. Of course, we do not really replace the vehicle stock overnight, or all in one year; rather, we replace it gradually, as vehicles retire. But in the long run, the annualized cost of replacing the existing fleet gradually is the same as the annualized cost of replacing it all at once. If vehicles have a life of n years, and every year 1/ nth of the vehicle stock is replaced, then the cost, calculated today, of each future 1/ nth fleet replacement is an annualized cost stream equal to 1/ nth the annualized cost of replacing the entire stock. These yearly annualized cost streams accumulate for n years, at which point we will have turned
9
over the entire stock and will have accumulated n annualized cost streams each 1/ nth the annualized cost of replacing the entire stock all at once.
We may conclude, then, that in this analysis we have estimated either the immediate annualized cost of replacing the entire stock overnight, or the annualized cost in the long run of continuing to replace vehicles as they retire.
Our estimate of the annualized cost is developed in Tables 4 and 5. The estimate excludes sales taxes, but still includes income taxes and charges, such as CAFE fines, leveled on producers. These are deducted en masse later. Also, we deduct what essentially is a guess at producer surplus, allowing, as discussed in Report # 1, that producer surplus that accrues to foreign producers should not be deducted, because it is a cost to consumers in the U. S.
It also happens, happily, that this estimate of the vehicle stock is exactly equal to my best independent estimate of the actual vehicle stock in 1991. According to FHWA, 188.3 million vehicle registrations occurred in 1991 ( FHWA, 1993). However, the FHWA registration data double count vehicles that were registered twice in the same year ( say, in different states), and hence overestimate the number of vehicles in use. R. H. Polk provides a better estimate of vehicles in use, because it counts only vehicles registered as of July 1 of any year. For 1991, Polk estimates that there were 181.4 million registered vehicles ( Davis, 1995). However, neither FHWA nor Polk account for unregistered vehicles in use. If these are 4% of the total ( the rate in California, according to Marowtiz, 1991), then the Polk data imply a total in- use fleet ( registered plus unregistered) of 189 million -- exactly my estimate here.
Of course, this equality is partly fortuitous, because vehicles sales can fluctuate considerably from year to year. In fact, in 1991, vehicle sales were at their lowest in many years, in part because of the recession ( Moran, 1991). If one performed this same calculation with 1989 sales, one would overestimate the 1989 vehicle stock. In general, if the vehicle stock is growing, then current sales multiplied by current life will exceed the current stock.
Salvage value. The present worth of the salvage value of vehicles at the end of their lives should be deducted from the up- front replacement cost before it is annualized. Rather than do that, however, I assume that the salvage value at the end of the life is about equal to the disposal and dismantling cost, and hence ignore both the salvage value and the disposal cost
5.2.2 The cost of transactions involving used cars
The preceding calculation annualizes the replacement- cost of the motor- vehicle fleet, where the replacement cost per vehicle is the present retail cost per new vehicle. This first retail cost naturally includes all the costs of the first transaction between dealer and buyer. It does not, however, include the transactions costs of subsequent transfers of the vehicle. Consequently, we must estimate and add separately the cost of transactions involving used cars.
In the case of used- car transactions that involve a car dealer, the cost of the transaction is equal to the dealer’s margin. To estimate the dealer’s margin on all used-
10
car transactions, we can use the method that the BEA uses to estimate PCEs on used automobiles9: the total dealer margin is equal to total sales of used cars multiplied by the dealer margin as a percent of sales. This estimate ( with the deduction of producer surplus) is developed in Table 5- 6.
Not all used- car transactions involve a dealer. Individuals and perhaps businesses and governments can transact between themselves. In Report # 4, we make a rough estimate of the time cost of transactions between persons. However, we do not estimate the cost of used- car transactions, without a dealer, between businesses or governments.
Disposal cost. There also is a cost to the disposal transaction at the end of the vehicle’s life. However, as explained above, I assume that the cost of disposal is about equal to the salvage value of the vehicle. I thus let the salvage value approximately cancel the disposal cost, and treat neither explicitly.
5.2.3 Deduction for external replacement costs due to accidents
The cost of replacing a vehicle totaled in an accident that is an externality should be classified as an monetary external cost, not a private cost. This is handled here by deducting the cost of all accidental property damage, whether private or external ( section 5.10.4).
9The BEA’s estimate of PCEs on used automobiles -- which as noted in Table 5- 1 is equal to the dealer’s margin on automobiles purchased by individuals plus net transactions ( purchases less sales) between persons and other sectors -- is not the same as the dealer’s margin on all used- car transactions, because it is limited to transactions involving persons. That is, it does not include the dealer’s margin on used cars and trucks purchased by governments and businesses, which we wish to include, and it inappropriately ( from our standpoint) includes net transactions between persons and other sectors, which we do not wish to include because we are considering all used- car transactions.
11
5.3 THE COST OF MOTOR FUEL AND LUBRICATING OIL, EXCLUDING EXCISE AND SALES TAXES AND THE COST OF EXTRA FUEL USED BECAUSE OF TRAVEL DELAY
5.3.1 Model of the cost of motor fuel
The total cost of motor fuel is equal to the price of the fuel, excluding taxes and fees, multiplied by the quantity consumed, less the portion of the price- times- quantity revenues that is producer surplus ( accrued to U. S. producers) rather than resource cost.
In this analysis, we separate the total fuel cost into the portion that is an externality due to traffic congestion, and the remainder that is not an externality. Traffic congestion causes an externality of additional fuel consumption because the fuel economy of vehicles is less during congestion than during free flow. The cost of excess fuel consumed during congestion that would not have been consumed had traffic been free flowing is a monetary externality, estimated here but included with the monetary externalities of Report # 8. The cost of the remaining fuel is a private cost, estimated and included here.
Formally, we estimate the cost of fuel as follows:
FCt= FCi+ FCe
FCe= Ge⋅ Pe− PSe= Ge⋅ Pe⋅ 1− PSeGe⋅ Pe⎛ ⎝ ⎜ ⎞ ⎠ ⎟ Assume: Pe= Pa and PSeGe⋅ Pe= PStGt⋅ Paand let: PStGt⋅ Pa= PSF
Then: FCe= Ge⋅ Pa⋅ 1− PSF() and similarly: FCi= Gt− Ge()⋅ Pa⋅ 1− PSF() eq. [ 5- 2a, b]
where:
FCt = the total fuel cost ( 109 1991$)
FCe = the fuel- cost externality, due to traffic delay ( 109 1991$)
FCi = the private- sector ( internal) fuel cost ( 109 1991$)
Ge = the motor- fuel- consumption externality: excess fuel consumed due to traffic delay ( 109 gallons) ( estimated below and shown in Table 5- 7)
Gt = total motor- fuel consumption ( 109 gallons) ( Table 5- 7)
12
Pe = the price of the excess motor- fuel consumed due to traffic delay, excluding taxes and fees ($/ gallon)
Pa = the average price of all motor= fuel consumed, excluding taxes and fees ($/ gallon) ( estimated below and shown in Table 5- 7)
PSe = the domestic producer surplus associated with the excess price- times- quantity payments ( for the excess fuel consumed due to traffic delay) ( 109 1991$)
PSt = the total producer surplus associated with all price- times- quantity payments ( 109 1991$).
PSF = the average producer- surplus fraction ( estimated below and shown in Table 5- 7)
This method assumes that the marginal costs ( Pe) and marginal producer- surplus shares ( PSe/( Ge. Pe)) that pertain to the fuel- consumption externality are equal to the average costs ( Pa) and average producer- surplus shares ( PSF) that pertain to all fuel consumption.
Our estimate of the total fuel consumption, Gt, is shown in Table 5- 7, and is based ultimately on FHWA data on VMT and gallons consumed. How accurate are these FHWA data? The data on VMT are derived from traffic counts made by the states, and probably are as accurate as any VMT data could be. The gallonage data are “ based on reports from State motor- fuel tax agencies” ( Highway Statistics, annual). This might be a problem, especially in the case of diesel fuel, because it is likely that there is some cheating to avoid paying taxes10. ( Some researcher believe that 15- 20% of diesel fuel is illegally untaxed.) However, I cannot find any evidence that the FHWA’s estimates of gallons consumed underestimate true consumption, for any reason. In the first place, both FHWA and the States [ obviously] account for legally untaxed gallonage: the FHWA estimates the use of gasoline by public vehicles, and the states estimate consumption of “ special fuels” ( mainly diesel fuel) by vehicles that pay a mileage tax and hence are exempt from the gallonage tax. Second, my best independent estimate of total consumption of diesel fuel in 1987 actually is lower than the FHWA’s estimate ( Table 5- 8). Third, the EIA uses the FHWA data without adjustment in its ( the EIA’s) estimates of diesel- fuel consumption by end use sector ( EIA, Fuel Oil and Kerosene
10The FHWA’s estimates of volumes of motor gasoline reported by wholesale distributors to State motor- fuel tax agencies ( Highway Statistics, annual) are about 3% less than the amounts reported in the EIA’s census of sales of refiners and gas- plant operators ( form EIA- 782A, Petroleum Marketing Annual, annual) ( Hallquist, 1994). Hallquist ( 1994) believes that the FWHA estimates are lower in part because “ tax avoidance causes undercounting” in the FHWA data ( p. xvii), and in part because of double counting in the EIA form 782A estimates. Also, it appears to me that the FHWA estimates are lower ( by about 1%) because they exclude gasoline exported and gasoline used by the military. If ( say) one percentage point of the 3% difference is due to double- counting on EIA 782A, and another point is due to the exclusion of military use and exports from the FHWA but not the EIA data, then under- reporting due to tax avoidance is about 1%.
13
Sales 1991, 1992). I conclude, therefore, that the FHWA has not seriously underestimated consumption of diesel fuel by motor vehicles11.
We now need to derive or estimate the parameters Ge, Pa, and PSF.
5.3.2 Excess fuel consumption due to traffic delay ( parameter Ge)
We estimate Ge, the motor- fuel- consumption externality due to traffic delay, as the difference between the amount of fuel actually consumed during delay, and the amount that would have been consumed had traffic not been delayed:
Ge= Gd− Gnd= VMTdMPGd− VMTdMPGnd eq. [ 5- 3]
where:
Ge is as defined above
Gd = the amount of fuel consumed during any conditions of traffic delay ( i. e., any conditions other than free flow) ( 109 gallons)
Gnd = the amount of fuel that would have been consumed over the mileage subject to delay had traffic been completely free flowing ( 109 gallons)
VMTd = vehicle miles of travel subject to delay ( 109; estimated below as a fraction of total VMT)
MPGd = the fuel economy of traffic during conditions of delay ( miles/ gallon; expression derived below, equation 5- 5a)
MPGnd = the fuel economy that would have been obtained over the mileage subject to delay had traffic not been delayed ( miles/ gallon; expression derived below, equation 5- 5b)
11Similarly, estimates of gasoline consumption derived from the 1987 economic Censuses are less than the FHWA’s estimate of gasoline consumption in 1987. The FHWA estimates that highway vehicles used 109 billion gallons of gasoline in 1987 ( Highway Statistics 1987, 1988). Using data from the 1987 Census of Retail Trade, Miscellaneous Subjects, ( Bureau of the Census, 1990), and the 1987 Census of Retail Trade, Merchandise Line Sales ( Bureau of the Census, 1990) I estimate that retail establishments sold 86 billion gallons of gasoline in 1987. Bulk plants and bulk terminals sold 93 billion gallons of gasoline wholesale in 1987 ( Bureau of the Census, 1987 Census of Wholesale Trade, Subject Series, Miscellaneous Subjects, 1991). Although neither the Census of Retail Trade nor the Census of Wholesale Trade cover all gasoline end use ( because, on the one hand, some wholesale and service establishments, which are not covered in the Census of Retail Trade, sell to end users, and, on the other, not all gasoline passes through a wholesaler), they clearly cover the great bulk of it, and hence the significant shortfalls between the Census estimates and the FHWA estimate do not support the hypothesis that the FWHA seriously underestimates gasoline consumption.
Of course, it is possible that the FHWA estimates are accurate, but that still, a lot of diesel fuel is illegally untaxed. If this is true, and if in the future the amount of fuel illegally untaxed fuel declines, then user payments for the highways ( estimated in Report # 17 of this social- cost series) will increase, regardless of what happens to tax rates.
14
It will be useful to express the fuel economy and VMT parameters in terms of other quantities known or at least easier to estimate. First, we will derive workable expressions for MPGd, and MPGnd, by starting with the proposition that total gallons of fuel consumed equals the gallons consumed during conditions of delay, plus gallons consumed during conditions of no delay. Then we will substitute these expressions back into equation 5- 3.
Gt= Gd+ Gnd^= VMTdMPGd+ VMT− VMTdMPGnd^ eq. [ 5- 4]
where:
Gt, Gd, and MPGd are as defined above.
Gnd^ = the amount of fuel consumed under conditions of no delay ( free flow) ( 109 gallons)
MPGnd^ = the fuel economy of vehicles under conditions of no delay ( free flow)
VMT = total vehicle- miles of travel ( 109)
Note that Gnd^ in equation 5- 4 is not necessarily equal to Gnd in equation 5- 3, and that MPGnd^ in equation 5- 4 is not necessarily equal to MPGnd in equation 5- 3. Gnd^ and MPGnd^ pertain to VMT that at present is not subject to delay, whereas Gnd and MPGnd pertain to hypothetical free- flow conditions over mileage that at present actually is subject to delay. Generally, because VMT not subject to delay exceeds VMT that is subject to delay, Gnd^ will exceed Gnd. However, unless delay occurs disproportionately on one particular type of road ( say, limited- access highways rather than city streets), the fuel economy under actual ( present) free- flow conditions generally will be close to the fuel economy that would obtain over presently delayed VMT12. So, it probably is reasonable, and certainly is analytically convenient, to assume that MPGnd^ = MPGnd.
We now proceed as follows:
12Fuel economy is determined by the grade of the road, the wind speed, the condition of the pavement, traffic density, the maximum allowable speed, the number and nature of intersections, the characteristics of the vehicles, and other factors. Thus, if at present delay occurs mainly on steep, pot- holed roads with lots of intersections, the fuel economy that would obtain over these roads were the delay eliminated still would be relatively low -- lower, certainly, then the fuel economy obtained over the presently undelayed, flat, smooth, uninterrupted roads.
15
Let: MPGnd= k1⋅ MPGd and VMTd= k2⋅ VMTAssume: MPGnd^= MPGndThen we have: Gt= k2⋅ VMTMPGd+ VMT− k2⋅ VMTk1⋅ MPGdMPGd= k2⋅ VMTGt+ VMT− k2⋅ VMTk1⋅ Gt= k1⋅ k2⋅ VMT+ VMT− k2⋅ VMTk1⋅ Gt
MPGd= VMT⋅( k1⋅ k2+ 1− k2) k1⋅ GtMPGnd= VMT⋅( k1⋅ k2+ 1− k2) Gt eq. [ 5- 5a, b]
where:
k1 = the ratio of fuel economy if no delay ( for presently delayed miles) to fuel economy under delay ( see parameter k below)
k2 = the ratio of delayed VMT to total VMT ( derived below)
16
Next, we substitute the expressions for MPGd and MPGnd ( equations 5a and 5b) into the expression for Ge, from equation 5- 3:
Ge= VMTdMPGd− VMTdMPGnd= k2⋅ VMTVMT⋅( k1⋅ k2+ 1− k2) k1⋅ Gt− k2⋅ VMTVMT⋅( k1⋅ k2+ 1− k2) Gt= k2⋅ k1⋅ Gt− k2⋅ Gt( k1⋅ k2+ 1− k2) = Gt⋅ k1− 1k1− 1+ 1k2Let: k1− 1= k
Ge= Gt⋅ kk+ 1k2 eq. [ 5- 6]
where:
Gt, Ge, and k2 are as defined above
k = the fractional increase in fuel economy, over presently delayed miles, that would result were the delays eliminated.
Finally, we can express the parameter k3 in terms of other parameters that are easier to estimate:
k2= VMTdVMTVMTd= VHT×Fd×SdVMT= VHT×Fd×Sd+ VHT×1− Fd() ×R×Sdk2= VHT×FdVHT×Fd+ VHT×1− Fd() ×R
k2= 11+ 1Fd− 1⎛ ⎝ ⎞ ⎠ ×R eq. [ 5- 7]
where:
17
VMTd = vehicle miles of travel subject to delay ( not in final equation)
VHT = total vehicle hours of travel ( not in final equation)
Fd = fraction of total vehicle hours of travel subject to delay ( discussed below)
Sd = average vehicle speed during delay ( not in final equation)
R = ratio of average speed when not delayed to average speed during delay ( discussed below)
VMT = total vehicle miles of travel ( not in final equation)
Leaving us with our final expression for the excess fuel consumed ( Ge):
Ge= Gt⋅ kk+ 1+ 1Fd− 1⎛ ⎝ ⎞ ⎠ ⋅ R eq. [ 5- 8]
The parameters Fd and R are estimated as follows:
LDGAs, LDDAs
LDGTs, LDDTs
HDGVs, HDDVs
Use the values for “ Private vehicles, personal purposes, daily travel,” in Table 4- 1 of Report # 4
FdLDT= PHT1⋅ Fd1+ PHT2⋅ Fd2PHT1+ PHT2
FdLDT = the parameter Fd for light- duty gasoline and diesel- fuel trucks
PHT1 = person- hours of travel in LDTs as personal household vehicles ( 31% of total person- hours in “ Private vehicles, personal purposes, daily travel” in Table 4- 1; 31% based on data in Hu and Young, 1992)
PHT2 = person- hours of travel in “ Light- duty trucks, no paid drivers” in Table 4- 1
Fd1 = the parameter Fd for “ Private vehicles, personal purposes, daily travel,” in Table 4- 1
Fd2 = the parameter Fd for “ Light- duty trucks, no paid drivers” in Table 4- 1
The value RLDT is calculated analogously.
Use the values for “ Heavy- duty trucks, paid drivers” in Table 4- 1 of Report # 4
18
Note that when we partition total cost to its external and internal components, we will designate the “ high” cost case that which results in high external costs ( and hence low internal costs).
5.3.3 The pre- tax cost of gasoline and diesel fuel ( parameter Pa)
For our price- times- quantity estimate of cost, in this section, we need to know the average pre- tax price of gasoline and diesel fuel. In other sections of this report, we need to know the final retail price, including taxes. Now, the EIA reports the sales- weighted retail price of gasoline, but not diesel fuel. It also reports the pre- tax price of gasoline and diesel fuel at refinery- owned stations, but not the price at all stations. ( The price at refinery- owned stations probably is less than the price at all stations, because the refinery- owned stations sell to bulk customers, who customarily are charged less per unit than are smaller volume customers.) The data situation is thus:
gasoline
diesel
pre- tax price at all stations
estimate from pre- tax price at refinery- owned stations ( which is reported by EIA)
estimate from pre- tax price at refinery- owned stations ( which is reported by EIA)
retail price ( including taxes) at all stations
reported by EIA
estimate as pre- tax price, above, plus all taxes
These estimates of prices are derived in Table 5- 9. The pre- tax price of gasoline and diesel fuel at all stations is equal to the pre- tax price at refinery- owned stations multiplied by an adjustment factor, shown in Table 5- 9. The adjustment factor is estimated such that the factor multiplied by the pre- tax price of gasoline at refinery- owned stations, plus all estimated gasoline taxes, is equal to the retail gasoline price reported by the EIA. The retail price of diesel fuel is estimated as the pre- tax price at all stations ( which as mentioned is equal to the pre- tax price at refinery- owned stations multiplied by the adjustment factor) plus Federal and state excise taxes and state sales taxes.
The estimated adjustment factor of 1.08 ( Table 5- 9) means that, if my estimates of taxes are correct, and if the EIA’s estimate of the actual sales- weighted selling price is correct, then it must be that the true pre- tax sales- weighted price of gasoline at all stations is 8% higher than the pre- tax price of gasoline at refinery- owned stations. This is plausible, because as mentioned above refinery- owned stations sell some fuel to bulk customers.
Note that the estimate of the true pre- tax price of diesel fuel at all stations, and hence the estimate of the retail price of diesel fuel, uses the adjustment factor derived from the gasoline data. That is, I assume that the pre- tax price of diesel fuel at refinery- owned stations underestimates the price of diesel fuel at all stations by the same factor that the pre- tax price of gasoline at refinery- owned stations underestimates the pre- tax price of gasoline at all stations. That the adjustment factor for gasoline appears to be
19
absolutely constant over time ( as shown in Table 5- 9, it was identical in 1987, 1991, and 1992) suggests that there is a systematic difference between refinery- owned outlets and all outlets, and gives me confidence that the factor can be applied to diesel fuel.
5.3.4 Producer surplus associated with motor fuels ( parameter PSF)
Many oil firms own relatively low- cost oil reserves, and hence earn sizeable producer surplus. In order to estimate this surplus, we need to estimate the supply curve and subtract the area under it ( the resource cost) from total price- times- quantity revenues.
Leiby ( 1993) estimates the following nonlinear marginal cost function for oil supply:
P= a+ bc− Q
where:
P = the price of supplying quantity Q ($/ bbl)
a = the price below which nobody producer will supply the market
c = the upper bound on supplies ( the price asymptote)
b = shape parameter
Q = the quantity of oil supplied ( million barrels/ day)
Given this, producer surplus PS can be estimated as:
PS= P* ×Q*− a+ bc− Q⎛ ⎝ ⎜ ⎞ ⎠ 0Q* ∫ dQ= P* ×Q*− 0Q* aQ− b×lnc− Q() = P* ×Q*− aQ*− b×lnc− Q*()+ b×lnc()()= = P* ×Q*− aQ*+ b×lnc− Q*()− lnc()() = Q* ×P*− a()+ b×lnc− Q*()− lnc()()
We estimate this for U. S. production only, because any producer surplus in price- times- quantity payments from U. S. consumers to foreign producers is a real net loss of wealth to the U. S. Using Leiby’s ( 1993) parameter values, we estimate that for U. S. oil producers PS is about 40% of price- times- quantity receipts. However, since on the order of half of all motor- fuel may be assumed to be either imported or made from imported crude oil, we want to deduct the 40% PS from about half of total fuel
20
consumption, which means that in effect we can assume that about 20% of price- times- quantity payments for the crude oil used to make motor- fuels is PS accruing to U. S. producers.
This, however, gives us the producers surplus in the oil industry only. We still should estimate PS in the downstream refining and marketing industries. Presumably, though, the downstream producers earn less surplus than do the oil producers, because unlike the oil producers, the refiners and marketers all probably have similar cost structures. Considering this, and allowing for uncertainty in the estimates of the domestic PS surplus fraction of crude oil in all motor- fuel, we assume that 20% to 30% of the pre- tax retail cost of gasoline and diesel fuel is PS accruing to domestic producers.
5.3.5 The cost of automotive lubricants sold at retail
I estimate the cost of automotive lubricants on the basis of retail sales reported by the Bureau of the Census.
Automotive lubricants are sold in the retail sector ( SICs 52, 53, 54, 55, 58, 59), in the automotive service sector ( SIC 75), and elsewhere. However, in this analysis, lubricants sold by service establishments, such as repair or lube shops, are included with the cost of parts, supplies, maintenance etc., estimated on the basis of sales in SIC 75. Therefore, the relevant total sales of lubricants, not covered elsewhere in this analysis, are those in the retail sector, reported in the Bureau of the Census Merchandise Line Sales series and those not covered in either the SIC 5- or SIC 75 sales data. I estimate data for 1991 by interpolating between 1987 and 1992 data ( 109 dollars):
1987
1992
Sales of automotive lubricants ( merchandise line 730) in SICs 52, 53, 54, 55, 58, 59 ( Bureau of the Census, 1987 Census of Retail Trade, Merchandise Line Sales, 1990; 1992 Census of Retail Trade, Merchandise Line Sales, 1995)
3.02
3.50
Sales of automotive lubricants outside of SICs 5- and 75 ( my estimate, based on data in the NIPAs [ Key, 1994])
0.19
0.21
Assuming that 25% of this is producer surplus ( the same percentage assumed for gasoline), and interpolating linearly, the resulting cost is $ 2.7 billion in 1991.
5.4 PARTS, SUPPLIES, MAINTENANCE, REPAIR, CLEANING, STORAGE, RENTING, TOWING, ETC., EXCEPT EXTERNAL COSTS OF ACCIDENTS
Our estimate of the cost of parts, supplies, maintenance, repair, and so on consists of:
• The cost of automotive services, which comprises:
-- receipts for automotive services in SIC 75
21
-- receipts for automotive services in SIC 55
-- in- house m& r expenditures by fleets
• The cost of parts and supplies, which comprises:
-- receipts for sales of new and rebuilt parts and supplies in SIC 55
-- sales of used auto parts, mainly SIC 5015
-- expenditures on merchandise used for motor vehicles but not classified as automotive merchandise
• Deductions for:
-- services in SIC 75 unrelated to motor- vehicle use
-- receipts for parking ( which are estimated separately)
-- U. S. producer surplus
• An estimate of the annualized cost of long- lived repairs
Since the Census estimates of receipts exclude sales taxes, we do not have to make any adjustments for sales taxes. In the following sections we discuss each of these items.
Note that the resulting estimate will include the cost of repair and replacement due to vehicle accidents as well as costs not due to accidents. Because we estimate and separately classify and list repair costs of accidents, we must deduct from the total estimate here whatever accident costs we estimate separately, in order to avoid double counting. This deduction to avoid double counting is accomplished in the section on accident costs ( 5.10.4).
5.4.1 The cost of automotive services
We begin with revenues received in 1991 in SIC 75, automotive services. This SIC includes only automotive service industries: automotive rental and leasing ( SIC 751), automobile parking ( 752), automotive repair shops ( 753), and automotive services except repair ( 754). The last includes washing, emissions testing, inspecting and diagnosing, lubricating, towing, wrecking, tinting, and rustproofing. The Census classifies establishments and presents data according to the Standard Industrial Classification system ( SIC) of the Office of Management and Budget ( OMB, 1987), as follows:
Auto service provided by:
SIC grouping:
Sales data:
establishments that primarily provide auto services to the general public
Major industry group 75
published for the whole SIC, in the U. S. Census’ quinquennial Census of Service Industries and annual Service Annual Survey 22
new- car dealers, used- car dealers, auto parts stores, and gasoline stations
Part of major group 55
included under “ non merchandise receipts” in the quinquennial Census of Retail Trade, Merchandise Line Sales
businesses, government for their own fleets
auxiliary establishments
not available ( I estimate separately below)
This is a mutually exclusive and exhaustive categorization of automotive services. Let’s address each of these in turn.
i) SIC 75. In 1991, firms subject to the Federal income tax in SIC 75 received $ 71.5 billion in revenues ( Bureau of the Census, Service Annual Survey: 1994, 1996). Apparently, there were no tax exempt firms in SIC 75.
ii) SIC 55. Some retail firms, in SIC 55, also provide automotive services. In 1987, firms in SIC 55 received $ 22.24 billion for automotive services ( Table 17- 14 of Report # 17). ( I count as automotive service all “ nonmerchandise receipts” in SIC 55, except sales of “ parts installed in repair,” “ credit life insurance and financing commissions,” and “ miscellaneous merchandise”. I count parts installed in repair separately, in the next section.) This was 43.3% of the $ 51.423 billion received in SIC 75 in 1987 ( Bureau of the Census, 1987 Census of Service Industries, United States, 1989). Thus, I add 43.3% to the receipts, reported above, in SIC 75 in 1991.
iii) In- house work at business and government fleets. Some business and government fleets perform maintenance and repair in house. If an in- house maintenance and repair shop does not qualify as a separate establishment in the SIC, then the cost of the work at the shop will not be included in the receipts in SIC 75 or 55.
It is difficult to estimate the cost of maintenance and repair work done in- house at government and business fleets. In Table 10- 7 of Report # 10, I estimate that in SIC 4212, local trucking, the $/ gallon maintenance and repair cost excluding the cost of in- house labor13 is the same as the $/ gallon cost for personal automobiles. Assuming that LDTs in SIC 4212 should have the same $/ gallon maintenance and repair cost as do personal LDAs, the estimated equality might suggest that in- house expenditures in SIC 4212 are not significant. This, however, would be an incorrect assumption, because as estimated in Report # 4, people spend a lot of their personal time repairing and maintaining their cars. Thus, it is possible that there are significant in- house expenditures on maintenance and repair in SIC 4212.
On the assumption that the maintenance and repair cost per mile of travel for fleet vehicles with in- house maintenance and repair is the same as the maintenance and repair cost per mile for vehicles repaired outside, I estimate the expenditures for in- house maintenance and repair, Mih, as follows:
13The maintenance and repair expenditures reported by the Census for SIC 421 include only the amounts paid to other firms ( Bureau of the Census, Motor Freight and Transportation Warehousing Survey: 1993, 1995).
23
Let: Fih= MihMih+ MoThen: MRih= Fih1− Fih⋅ MRo eq. [ 5- 9a]
Estimate Fih as: Fih= FMih⋅ VMTv⋅ FVMTvvΣVMTt eq. [ 5- 9b]
where:
MRih = expenditures for in- house maintenance and repair ($)
Fih = of total maintenance and repair expenditures, the fraction that is in- house
MRo = expenditures for outside maintenance and repair ($; revenues in SICs 75 and 55, as discussed above)
FMih = of total maintenance and repair expenditures at fleets, the fraction that is in- house ( in the absence of any data, I assume 25% to 50%)
VMTv = total vehicle miles of travel in category v of Table 4- 1 ( private vehicles for personal purposes, private vehicles for business purposes, etc.)
FVMTv = of VMT in each category v, the fraction that is by fleet vehicles ( I assume 1.00 for all buses, government vehicles, and private heavy- duty vehicles, 0.00 for private vehicles used for personal purposes, 0.80 for private LDTs used for business purposes, and, on = the basis of data in Miaou et al. ( 1992), 0.60 for VMT by private LDAs used for business purposes)
VMTt = total VMT in 1991 ( Table 4- 1).
iv) Personal time spent maintaining and repairing vehicles. This I classify as a personal nonmonetary cost, and estimate in Report # 4.
v) One more item. Note that all motor- vehicle damage to buildings that is paid for by the responsible party is included below, under “ Accident costs paid for by responsible party, but not through automobile insurance...” That is, I distinguish properly priced vehicular damage from properly priced damage to buildings in part because most analysts consider the former but not the latter, which admittedly is very small.
24
5.4.2 The cost of parts and supplies
New and rebuilt parts and supplies. Next, we must add in receipts for parts, supplies, tires, accessories, and the like. In Table 17- 15 of Report # 17, we estimate that receipts for automotive parts and supplies, including parts installed in repair in SIC 55, were $ 61.7 billion ( excluding sales taxes) in 1991.
Used parts and supplies. According to the Census’ Classification Manual ( Bureau of the Census, 1992), the auto and home supply stores of SIC 553 sell new and rebuilt -- but not used -- automobile parts and accessories. In support of this, the Census Merchandise Line Sales ( Bureau of the Census, 1995), shows $ 11.5 billion in sales of new and rebuilt parts in SIC 55, and only $ 68 million in sales of used parts. In the Census system, sales of used parts are classified as “ wholesale,” and occur mainly in SIC 5015, “ motor vehicle parts, used”. In 1992, sales of “ used automotive parts, accessories, and equipment” ( commodity line 0240) were $ 3.571 billion ( Bureau of the Census, 1992 Census of Wholesale Trade, Subject Series, Commodity Line Sales, United States, 1995). I assume that in 1991 sales were 2% less.
Parts and supplies sold in non- auto stores and not classified as automotive merchandise. Finally, I account for expenditures on items, such as all- purpose tools, that are used for motor vehicles but or not sold in automotive stores or classified as automotive merchandise. I assume that expenditures on such items are 1% to 2% of the expenditures on new automotive parts classified as such.
5.4.3 Deductions
Services unrelated to motor- vehicle use. We deduct from total receipts in SIC 75 those that were for services unrelated to motor- vehicle use. Naturally, this is a very small fraction of the total. The Bureau of the Census 1992 Census of Service Industries, Subject Series, Sources of Receipts or Revenue ( 1996) breakdowns receipts in SIC 75 by source. The categories “ all other receipts from customers” and “ all other receipts” appear to comprise mainly non- motor- vehicle services, because all major motor- vehicle services, as well as a category “ all other motor vehicle services,” are listed separately. In 1992, “ all other receipts from customers” and “ all other receipts” were 1.7% of total receipts in SIC 75. I assume, therefore, that 1.5% of total receipts in SIC 75 were unrelated to motor- vehicle use.
Parking. We also deduct receipts in SIC 752, parking, because we count those separately in this report ( section 5.8).
Producer surplus. Finally, we deduct the producer surplus that accrues to U. S. producers. I assume that most firms in this industry have a similar cost structure, and hence that producers surplus is relatively small. I assume 5% to 10% for all producers. However, foreign producers of automotive parts earn about 1/ 3 of all of the revenues earned from the sale of automotive parts in the U. S. ( International Trade Administration, 1995). This foreign producer surplus is a real cost to the U. S. Therefore, I assume that the producer surplus that accrues to U. S. producers is 3% to 8% of total revenues.
25
5.4.4 Estimating the annualized cost of long- lived repairs
The cost of any long- lived repairs -- i. e., the cost of replacing any major, long- lived components of the vehicle, but not the whole vehicle -- must be annualized, just as the cost of the vehicle fleet itself is annualized. For example, the cost of replacing an engine or transmission probably should be annualized over the life of the vehicle. To annualize the cost of replacing long- lived components, I first estimate annual expenditures for major, long- lived capital replacement ( as distinguished from expenditures for short- lived, operational repairs), and then annualize the fleetwide expenditures over their life14. With this method, the total cost of maintenance and repair is equal to annual expenditures on short- lived maintenance and repair plus the annualized fleetwide cost of long- lived repair and replacement.
The estimate of the annualized fleetwide cost of long- lived repairs thus begins with an estimate of annual expenditures on long- lived repairs. Here, I distinguish four kinds of expenditures:
i) replace vehicles damaged in motor- vehicle accidents
ii) replace major long- lived components of the vehicle damaged in motor- vehicle accidents
iii) replace vehicles worn out at the end of their of normal life
iv) replace major long- lived components of the vehicle worn out at the end of their normal life
I distinguish the replacement of the vehicle ( items i and ii) from the replacement of major components ( items ii and iv) because the former is part of the annualized cost of the vehicle fleet, already estimated as annualized cost in section 5.2.1, whereas the latter is part of the annualized cost of maintenance and repair, to be estimated in this section. I distinguish the replacement of parts damaged in accidents from the replacement of parts worn out at the end of their normal life because I estimate the annualized cost of property damage in accidents as part of my overall estimate of the cost of motor- vehicle accidents. However, I first estimate accident and non- accident component replacement costs ( ii and iv) together in this subsection, and then make a separate estimate of the accident- related components in Report # 19 and section 5.10.4 of this report.
14 There are two differences between an annual expenditure and the annualized fleetwide cost. The annual expenditure applies to only a portion of the fleet ( because only a portion incurs the cost every year), and is capital value only, with no interest ( opportunity- cost- of- money) component. The fleetwide annualized cost is the accumulated capital value of all replacements over the entire fleet over all years, converted to an equivalent annual stream that includes an interest component. On the assumption that every year the annual replacement expenditure is made on 1/ L of the fleet, where L is the life of the replacement in the annualization calculation, then the annualized cost is equal to the annual expenditure multiplied by a factor that accounts for the opportunity cost of money.
26
In this subsection, we are interested in items ii) and iv), the expenditures to replace major components damaged in accidents or worn out at the end of their normal life. First we estimate the annual expenditures, and then we estimate the annualized fleetwide cost. We estimate the annual expenditures on replacing long- lived components as follows:
COM= CAPA⋅ COMAF⋅ 1+ COMWF() eq. [ 5- 10]
where:
COM = annual expenditures to replace major long- lived components ( 109 $/ year)
CAPA = annual expenditures to replace all long- lived capital, including complete vehicles, damaged in motor- vehicle accidents ( 109 $/ year) ( section 5.10.4 and Report # 19)
COMAF = of expenditures to replace all capital damaged in motor- vehicle accidents, the fraction that is for replacing long- lived vehicle components ( e. g., transmissions) rather than complete vehicles themselves ( I assume 0.30 to 0.40)
COMWF = expenditures to replace major components worn out at the end of their life, as a fraction of expenditures to replace major components damaged in accidents ( I assume 1.5 to 2.0)
This method of relating the total capital- replacement expenditure to the expenditure to replace capital damaged in accidents ensures that the estimates of i), ii), iii), and iv) are consistent.
With these assumptions and data, I estimate that COM in equation 5- 10 is about $ 16 to $ 26 billion. By comparison, in 1992, some $ 30 billion worth of motor- vehicle parts were sold in the retail trade sector ( including parts installed in repair) ( Bureau of the Census, 1992 Census of Retail Trade, Merchandise Line Sales, 1995), and some $ 45 billion worth of repair and maintenance services were sold by automotive service establishments ( Bureau of the Census, 1992 Census of Service Industries, Subject Series, Sources of Receipts or Revenue , 1996).
The annualized fleetwide cost is estimated given these annual expenditures. The annualization method annualizes the value of the entire “ stock” of long- lived capital replacements, for the entire fleet, over the average life of the replacement, using the standard amortization formula ( equation 5- 1). The capital value of the stock of long- lived replacements for the entire vehicle fleet ( parameter I in equation 5- 1) is assumed to be equal to annual expenditures multiplied by the average life L of the replacement, on the assumption that the yearly annual expenditure replaces 1/ L of the fleetwide stock. The average life is assumed to be the average life of the entire vehicle fleet ( Table 5- 4), and the relevant interest rate ( parameter i in equation 5- 1) is assumed to be the fleetwide average used to annualize the cost of the vehicle fleet ( Table 5- 4).
27
5.4.5 Allocation to six classes of vehicles
The costs estimated in the preceding two sections are for the entire vehicle fleet. Unfortunately, the available data do not make it easy to allocate these costs for different vehicle classes. As a rough guide, one can use the maintenance and repair allocation factors of Table 10- 3 in Report # 10 of this social- cost series. These factors are estimated on the basis of personal- consumption expenditures on maintenance and repair of automobiles, and purchased maintenance and repair of trucks in SIC 421. ( Note that the maintenance and repair costs of Table 10- 3 are defined more narrowly than are parts, supplies, maintenance, repair, and so on here, and hence come to much lower grand total.)
5.5 AUTOMOBILE INSURANCE: ADMINISTRATIVE AND MANAGEMENT COSTS, AND PROFIT
5.5.1 An estimate of the cost
The actual resource cost of automobile insurance is the administrative and management cost of providing the insurance service. There are at least four kinds of insurance to consider:
i) Insurance provided by private insurance companies
ii) “ self- insurance” by government
iii) self- insurance by private companies
iv) private insurance by posted bond
Insurance provided by insurance companies. A reasonable estimate of the administrative and management cost of automobile insurance companies is the total underwriting and claims adjustment expenses. Data on these expenses are available.
The primary source of data on premiums and expenses in the insurance industry is A. M. Best’s Aggregates and Averages, Property- Casualty. ( The Bureau of Economic Analysis uses Best’s data in its National Income Product Accounts.) Table 5- 10 shows Best’s ( 1992) estimates of premiums and expenses for liability insurance and collision damage insurance for private passenger vehicles and commercial vehicles in 1991. The total expenses were $ 35 billion.
Because this estimate of cost is based on company- reported expenses, rather than price- times- quantity revenues, there is no need to deduct producer surplus.
Self insurance by government and private companies. Although governments presumably are large enough that they can afford to pay automobile accident costs as they go and so do without auto insurance altogether, they still will incur some insurance- like administrative and management costs when they process payments and claims. Similarly, some large commercial fleets, such as those at universities, car rental
28
companies, and utility companies, are self insured, but still will incur some insurance- like administrative and management costs.
I estimate the administrative and management costs of motor- vehicle self insurance on the basis of travel by self- insured vehicles relative to travel by other insured vehicles, and the administrative and management cost of self- insurance, per VMT, relative to the administrative and management cost of other insured vehicles, per VMT.
In 1991, the VMT of government vehicles was about 1.9% of total VMT, or probably around 2.2% of VMT by all insured vehicles ( Report # 4). According to the EIA ( 1996), in the early 1990s there were 10.5 to 12.3 million vehicles in non- governmental fleets of 10 or more vehicles, including 1.1 million in utility fleets, 0.140 million taxis, and 1.75 million rental vehicles. If one- quarter of the 11 or so million vehicles in large non- governmental fleets were self insured, and if self- insured vehicles had 1.5 times the VMT/ vehicle of other vehicles, then VMT by self insured non- government fleet vehicles was about 2.2% of VMT by all vehicles ( based on 190 million vehicles), or about 2.6% of VMT by all insured vehicles.
Thus, I estimate that VMT by self- insured government and non- government fleet vehicles was about 4.8% of VMT by all insured vehicles, or 5% of VMT by all vehicles insured by a motor- vehicle insurance company.
Presumably, the administrative and management cost of self- insurance, per VMT of travel by self- insured vehicles, is less than the administrative and management cost of private motor- vehicle insurance companies, per VMT of travel by vehicles insured by a motor- vehicle insurance company. The self- insured do not incur the sizable brokerage and commission expenses of motor- vehicle insurance companies, and do not have to write and administer policies. Also, they probably have lower costs of claims adjustment, and perhaps even lower general overhead costs ( because of shared building costs, for example). I will assume that the administrative and management cost per VMT for the self insured is one- half the cost for those insured by a motor- vehicle insurance company. With this assumption, the administration and management costs of self insurance are 2.5% of the of the actual insurance administration and management costs of automobile insurance companies.
Private insurance by posted bond. In at least some states, it is permissible to post a bond as automobile insurance. ( In California, the minimum amount is $ 35,000.) Because these bonds can earn interest at normal market rates, and do not require the administrative services of an insurance company, they have essentially no cost. In any case, it is likely that very few vehicles are insured by bond. For example, in California in 1989, only 126 personal passenger vehicles were insured by cash bond ( Marowitz, 1991)
5.5.2 Our estimate vs. the “ net premiums paid by persons” in the NIPA
Our estimate of the administrative and management cost of providing motor- vehicle insurance is not the same as the BEA’s estimate of net personal consumption expenditures on motor- vehicle insurance. In its estimate of PCEs in the NIPA, the BEA uses the A. M. Best data to calculate what it calls the “ net insurance premium” paid by
29
persons. The net insurance premium is the difference between total premiums paid out by persons and claim reimbursements received back by persons. The BEA calculates this as follows:
i) net premiums earned for private passenger liability insurance and private passenger collision damage insurance ( Table 5- 10), less
ii) losses incurred for same ( Table 5- 10), less
iii) dividends paid for same ( which are a tiny amount, and not shown in Table 5- 10), less
iv) the small portion of “ private passenger” insurance, as defined by Best, that is written for businesses rather than persons ( Key, 1994).
Thus, the BEA estimates the net personal expenditure on automobile insurance ($ 22.7 billion in 1991), not the cost of running the automobile insurance industry. Compared with our estimate, they exclude certain kinds of costs, and of course all costs of insurance for commercial vehicles.
5.5.3 Are automobile insurance prices optimal?
Although automobile insurance is provided in a reasonably competitive market, insurance prices are not necessarily optimal. The economic efficiency of the present insurance system perhaps can be improved. How much the system can be improved depends on how costly it is to get accurate, detailed information about people, vehicles and trips, and to administer a detailed, sophisticated pricing scheme. Ideally, insurance -- or any charge for expected damage inflicted on others -- would be a function of the number of miles actually driven ( if you did not drive, you would be charged for expected damages), the time and location of the trip, the route taken, the characteristics of the road, expected traffic conditions, the up- to- the- minute characteristics of the driver and vehicle, and other factors. ( See Edlin [ 2002[ for a discussion and analysis of related issues.) In this ideal world the driver also would be able choose at any time to purchase any type of insurance against damage to herself and her property. In the real world, however, it is too costly to set and enforce prices based on all of the determinants of expected damage cost, and so prices are based on a few key determinants, such as the age and marital status of the driver, distance from home to work, and home location. Any simplified system will omit some important determinants of expected damage. The current system, for example, does not charge per mile of actual driving. By contrast, a scheme to add a universal liability charge to the price of gasoline ( see Tobias, 1993; and the Quad Report, 1993) would have the great advantage of making the expected- damage premium a continuous real- time function of the amount of travel, but the considerable disadvantage of failing to distinguish drivers according to the expected riskiness of their behavior.
30
5.7 PRICED PRIVATE COMMERCIAL AND RESIDENTIAL PARKING, EXCLUDING THE PARKING TAX
Although the vast majority of parking is unpriced ( see Report # 6), motor- vehicle users do pay several billion dollars per year to private parking operators. These price- times- quantity payments, less taxes and producer surplus, are the resource cost of priced private- sector parking in the U. S.
There are in principle three kinds of priced private parking to consider in this report:
• priced private on- street parking
• priced private off- street residential parking
• priced private off- street commercial ( nonresidential) parking
I address each of these in turn. ( The cost of unpriced or bundled private parking, such as an attached 2- car garage or free parking at a shopping center, is estimated in Report # 6, and the cost of all public ( municipal and institutional) parking is estimated in Report # 7.
5.7.1 Priced private on- street parking
There may be some priced parking spaces on privately owned streets ( for example, on streets in a gated community), but the total amount of such parking must be insignificant. I assume that the cost of parking in this category is zero. ( Alternatively, one can assume that the cost of this parking is included already in the estimates of the costs of private roads, which estimates are broad and loosely defined enough to include any on- street private parking. See Report # 6.)
5.7.2 Priced private off- street residential parking
As shown in the notes to Table 5- 1, consumers reported spending some $ 200 million on residential parking in 1991. Before I count this expenditure as a separate cost of privately owned parking, however, I must be sure that it does not double count other parking costs estimated in this report, to wit:
31
1). Is priced residential parking ( Table 5- 1) already counted in this analysis as private, off- street, unpriced residential parking? Most likely not: in Report # 6, the cost of private, off- street, bundled residential parking is estimated as the average cost per space multiplied by the quantity of spaces, and the estimate of quantity specifically excludes parking spaces that are not included with the house or in the rent.
2). Are the payments for residential parking already counted as receipts to commercial parking operators ( estimated below)? Presumably not: those who charge for residential parking probably are not parking establishments as defined by the Census classification, but rather just property owners who charge for parking separately rather than include it in the rent or ownership fee.
3). Are the payments for residential parking already counted as parking or road expenditures by government? In Report # 7, we estimate the cost of public parking on the basis of Census estimates of government expenditures on parking. These government expenditures are for the provision, construction, maintenance, and operation of local government parking facilities -- public parking lots and garages, and parking meters on- street and in lots -- operated on a commercial basis ( Bureau of the Census, Classification Manual, 1992) . They do not include expenditures for the enforcement of parking regulations, or for parking facilities connected to a specific type of facility, such as a sports stadium ( counted as an expenditures for the specific type of facility) ( Bureau of the Census, Classification Manual, 1992). Thus, unless local governments own and operate commercial residential parking facilities -- and I assume that they don’t -- the payments reported in Table 5- 1 for residential parking are not counted in Report # 7.
However, any consumer expenditures for on- street parking permits will double- count the cost of streets, which is estimated in full in Report # 7. I assume that any such double counting is minor, and ignore it.
It appears, then, that I may count most of the $ 200 million expenditure on residential parking ( less any taxes, which I assume to be zero, and less any producer surplus, as estimated below) as an additional cost.
5.7.3 Priced private off- street commercial parking
The cost of priced private off- street commercial parking is estimated as total revenues to commercial parking operators in SIC 752 ( Bureau of the Census, Service Annual Survey: 1994, 1996), less my estimate of producer surplus.
The Census estimate excludes revenue from parking lots and garages that are owned and operated by municipalities, from parking lots that are part of another business ( mainly airports, hospitals, restaurants, and universities), and from facilities that provide long- term or dead storage of automobiles ( McKenzie, 1993; Bureau of the
32
Census, 1992 Census of Service Industries, 1994). I consider all of this excluded parking to be municipal and institutional parking, and estimate the cost in Report # 715.
As mentioned above, I have assumed that the revenues to commercial parking operators, as reported to the Census, do not include any payments from persons for residential parking. I also assume that any potential double counting or undercounting of municipal parking costs also is small16.
5.7.4 Total cost of private commercial and residential parking
I thus estimate the total cost as follows:
Payments for on- street private parking
0.00
Payments for off- street private residential parking
0.20
Parking revenues received by commercial parking facilities in 1991 ( local taxes excluded) ( 109 $)
3.305
My estimate of the fraction that is producer surplus
0.10
Estimated cost of priced private commercial off- street parking ( 109 $)
3.2
5.8 TRAVEL TIME, EXCLUDING TRAVEL DELAY IMPOSED BY OTHERS, THAT DISPLACES PAID WORK
5.8.1 Background
The value of the time that people spend in their cars and trucks is the single largest item in my cost accounting. In this study, we estimate that all travel time in motor vehicles ( including compensation of professional drivers) is worth roughly one trillion dollars annually.
In general, the cost of any travel time, whether monetary or nonmonetary, personal or external, can be estimated simply as the amount of travel time, in hours,
15Some institutional parking, such as that provided by private universities, arguably should be classified as private parking, and ( if priced) included in this report. However, the amounts involved are relatively small, and the distinction in this case between public and private is relatively unimportant.
16The municipal parking excluded here is not quite the same as the municipal parking included in Report # 7. The Service Annual Survey estimates of revenues to “ commercial” operators include revenue from municipally owned but privately run facilities if the private operator provides the management and operating staff, but not if the private company provides only the management staff ( McKenzie, 1993). Given that in its Government Finance s series, the Census reports local government expenditures for parking facilities, we may conclude that neither the Service Annual Survey estimates of private parking revenues nor the Government Finance estimates of public parking expenditures cover the cost of private management at publicly owned and operated facilities. It also might be the case that the public expenditures for “ ownership” are in essence double counted in the revenues received by facilities publicly owned but privately run.
33
multiplied by the cost per hour of travel. Total travel time can be estimated in a straightforward manner from data on travel times or data on average speeds and distances ( see Report # 4). It is not so straightforward, however, to separate the externality of travel delay from the total travel time ( see the discussion in Report # 4 and Report # 9). And the cost per hour of travel time is considerably more difficult yet to define and measure.
In this section of this report, I estimate the value of travel time ( excluding travel delay) that displaces paid work, and the cost of driver time in light- duty and heavy- duty commercial trucks. The value of travel time, excluding travel delay, that displaces unpaid activities, is estimated in Report # 4. External costs of travel delay are included with the items estimated in Report # 8 and Report # 9, but actually are detailed in Report # 4.
5.8.2 The cost per hour of travel time: concepts.
We may define the cost of travel time as the social willingness to pay ( WTP) to have the travel time reduced to zero, all else ( including access) equal. In principle, this cost, or social WTP, has two components: an opportunity- cost component, and a hedonic component ( Hensher, 1997).
The opportunity cost is the value of activities foregone while in the car. Analytically, it is useful to distinguish monetary, or paid activities foregone, from nonmonetary, or unpaid activities foregone, because the dollar value of the paid activity is explicit, whereas the dollar value of the unpaid activity has to be estimated by non- market valuation or indirect market methods. Note that, if one is able to do in the car precisely what one would do were travel time reduced to zero, then there is no opportunity cost. Because the magnitude of the opportunity cost depends precisely on what is being foregone, it will vary considerably across individuals and trips. For simplicity, I will consider only two general kinds of foregone activities: leisure, or unpaid activities, and paid productive work. I will estimate the value of both with respect to the individual’s income.
The hedonic cost is the pure utility or disutility of the motoring experience itself. The hedonic cost is determined by several factors, including comfort, safety, privacy, available space, amenities, and the amount of effort and attention required to control or in general worry about a vehicle. However, because the hedonic cost is non- monetary, I include the entire amount with our estimates of non- monetary time costs, in Reports # 9 and # 4. Here, I estimate only the monetary opportunity cost of travel time.
See Report # 4 for further discussion.
5.8.3 Categories of travel, by type of vehicle, according to the data.
Because the cost per hour depends on the type of trip and the income of the traveler, I estimate cost per hour and travel time for several different kinds of trips and
34
trip- makers. In the first instance, I distinguish travel in the following general categories17:
• Private vehicles, for personal purposes
-- daily travel ( LDAs, LDTs)
-- long trips ( LDAs, LDTs)
• Private vehicles, for business purposes
-- LDAs ( without paid drivers)
-- LDTs, without paid drivers
-- LDTs, with paid drivers
-- HDTs ( with paid drivers)
• Buses
-- intercity and transit buses
-- school buses
• Public ( government) vehicles
-- federal civilian vehicles ( LDAs, LDTs, HDTs)
-- federal military vehicles ( LDAs, LDTs, HDTs)
-- state and local civilian vehicles ( LDAs, LDTs, HDTs)
-- state and local police vehicles
Within each travel category, I estimate the portion of the total travel time that is due to delay ( an external cost), and the portion that is not, and the portion of travel that displaces paid work, and the portion that displaces unpaid activities. The portion that is not due to delay and that displaces paid work is a monetary non- external cost, and is estimated next.
5.8.4 Estimating the cost
In each vehicle travel category, the monetary time cost of travel, excluding delay, is calculated simply as the total travel time, less person- hours of delay ( which are externalities, and treated in Reports 8 and 9), multiplied by the fraction of travel time that displaces monetary ( paid) activities rather than unpaid activities, and by the cost per hour of the foregone monetary activities:
TTCim= PHT− PHTd()⋅ 1Oc+ 1− 1Oc⎛ ⎝ ⎞ ⎠ ⋅ Pa⎛ ⎝ ⎞ ⎠ ⋅ Fm, dr⋅ Cm eq. [ 5- 11]
where:
17Hensher et al. ( 1990) distinguish four kinds of trips: 1) private commuting to work in household vehicles; ii) commuting to work in company- supplied vehicles; iii) travel as a part of work; and iv) non- work related personal travel. They distinguished between commuters using private vehicles and commuters using company vehicles because the latter have a higher income than the former, and are willing to pay a higher percentage of that income to save time.
35
TTCim = the internal, monetary travel- time cost ( 109 1991$)
PHT = total person- hours of travel time ( 109 person- hours of travel; Table 4- 1, Report # 4)
PHTd = person- hours of delay ( the travel- time externality) ( 109 person- hours of delay; Table 4- 1, Report # 4)
Oc = average vehicle occupancy ( persons/ vehicle; Table 4- 1, Report # 4)
Fm, dr = the fraction of travel time that displaces monetary ( paid) activities rather than unpaid activities, for drivers ( Table 4- 1, Report # 4)
Pa = the ratio of parameter Fm for passengers to Fm for drivers ( Fm, pa/ Fm, dr; Table 4- 1, Report # 4J)
Cm = the cost of the foregone monetary ( paid) activities ($/ person- hour; discussed below; shown in Table 4- 1, Report # 4)
5.8.5 The cost of foregone monetary activity ( parameter Cm).
In Report # 4, I assume that Fm, dr is equal to zero for three of the vehicle travel categories: daily travel in private vehicles for personal purposes; long trips in private vehicles for personal purposes, and travel in school buses. Thus, for these three travel categories, there is no need to estimate Cm, the monetary cost per hour. In the following sections, then, I will estimate Cm for the remaining categories.
Private LDAs and LDTs, without paid drivers, used for business purposes, and government vehicles: concepts. As Hensher et al. ( 1990) note, “ the value to the community of an employee spending less time traveling and more time in productive work is... approximately equal to the full wage rate” ( p. 154), which in their analysis is the pre- tax salary plus 34% for benefits and other compensation. I will use as an approximation of the value of foregone productivity during business or government travel the present average hourly compensation rate in private industry or in the public sector.
Table 5- 11 shows my estimates of average compensation rates by SIC classification. The travel time costs of Table 4- 1 are taken from the compensation rates of Table 5- 11, as follows:
Table 4- 1 category:
Private LDAs and LDTs without paid drivers, used for business
Federal civilian vehicles
Federal military vehicles
State and local civilian vehicles
Table 5- 11 value, low:
Private industry
Government: federal non- military
Government: federal military
Government: state and local
Table 5- 11 value, high:
Finance, insurance, real estate
Government: federal non- military
Government: federal military
Government: state and local
Note that these average compensation rates are but approximations of the value of the foregone productivity, because there is no reason to believe that the productivity that actually is foregone as a result of business or government travel is the same as the
36
“ average” productivity represented by the average compensation rate. In the first place, it may be that the business people who travel a lot tend to be less productive per hour ( when they are not traveling) than is the average private- sector employee. Ideally, in order to estimate the cost of time in business and government travel, I would make a detailed list of occupations, and get data on the amount of employee travel and the specific compensation rate in each type of occupation. Unfortunately, neither travel times nor full compensation rates are known for specific occupations, and so instead I estimate travel time and compensation rates for the broad categories shown in Table 4- 1.
Beyond that, even if in every business travel time is the same fraction of total work time, the value of any productivity foregone by travel still is not be equal to the average compensation rate, because the work that actually is foregone at the margin is not necessarily of the same type and value as that done on average. Indeed, if marginal productivity is not constant, and is a function of the amount of work time, then one can presume that productivity foregone during travel generally is of lower value than is the average productivity.
Nevertheless, I ignore these complications, and use average hourly compensation rates as shown. I base the estimate on the full hourly rate of employee compensation -- gross wages and salaries, tips, bonuses, benefits, and employer- paid taxes ( about 20% higher than gross wages and salaries) -- and not after- tax take- home pay, because that is the full cost of the employee to the employer, and in principle equals the marginal productivity of the employee ( Button, 1993; Hensher et al. 1990).
Private LDAs and LDTs, without paid drivers, used for business purposes, and government vehicles: estimates. My estimates of full hourly compensation, shown in Table 5- 11, are derived from data from the National Income Product Accounts of the U. S., for 1990 ( NIPA). The NIPA show total employee wages, total compensation, and total hours in industries classified according to the Standard Industrial Classification SIC) ( Survey of Current Business, July 1992). Table 5- 11 shows data from the NIPA for several SIC categories relevant to this analysis: all employment; all private industry; transportation and utilities; trucking and warehousing; finance, insurance, and real estate; services; private household services; federal civilian, federal military, and state and local government. I have included the full compensation in private- household services for comparison with my estimate of the value of personal travel time. I have included the full compensation rate in finance, insurance, and real estate as an alternative ( high- cost) measure of the value of business- travel time, on the assumption that employees in those industries travel a lot.
Table 5- 11 compares wage and compensation data from the NIPA with data from the BLS’s News, “ Employer Costs for Employee Compensation”, the BLS’s ES- 202Employment and Wages Annual Averages, and the BLS’s Current Population Survey ( CPS). Of the three BLS data series, only the “ Employer Costs for Employee Compensation” reports full employee compensation as well as employee wages. The compensation data in the NIPA are preferable to those from the BLS News, “ Employer Costs for Employee Compensation,” because the measure of total compensation in the
37
NIPA appears to be more comprehensive than the measure in the BLS. For example, it appears that the NIPA counts as part of “ wages” the cash value of lodging and meals, items which the BLS’ “ Employee Costs for Employee Compensation” News apparently does not count at all, as a wage or a benefit. Perhaps in part because of this, the average hourly compensation rate reported in the NIPA is higher than the hourly rage reported in the BLS’s “ Employee Costs for Employee Compensation” News ( Table 5- 11).
Some of the NIPA data are derived from the ES- 202 data collected by the BLS ( Employment and Wages Annual Averages 1990, 1991). ( See the BLS Handbook of Methods, 1992, for more information.) As shown in Table 5- 11, the NIPA data generally agree with the BLS ES- 202 data18. However, with one important exception, the NIPA data do not agree well with BLS data reported by occupation, from the CPS ( Table 5- 11, last column). ( They do not agree because “ wages” in the NIPA are defined differently than are “ earnings” in the BLS occupation data, and the SIC categories of the NIPA cover different workers than do the occupation categories of the BLS.) The important exception is that NIPA- reported average wages for trucking and warehousing are nearly the same as BLS- reported average weekly earnings for transportation and material moving occupations ( Table 5- 11). This agreement is important because, as I explain next, I use the occupational earnings data to estimate the cost per hour of commercial truck driving.
For more details on the data of Table 5- 11, see the Appendix to this report.
LDTs and HDTs with paid drivers The cost of an hour of a truck- driver’s time should be analyzed separately from the cost of an hour of a business traveler’s time, because the truck driver produces driving, which is valued directly by the driver’s compensation rate. That is, the full compensation paid truck drivers is a good, direct estimate of the social cost of an hour of a truck- driver’s time.
The cost of truck driving is the social value of whatever else the drivers would do were they not driving. At the margin, the social value of the next best productive alternative is equal to the compensation actually paid the truck drivers. That is, the compensation actually paid the drivers is the value of the driver’s next best opportunities.
There are no data on the full hourly compensation rate for truck drivers specifically. However, the Bureau of Labor Statistics does report the 1990 average weekly earnings of drivers of light- duty trucks, and the average weekly earnings of drivers of heavy- duty trucks ( Bureau of Labor Statistics, unpublished tabulations, 1993). I can derive an estimate of the of hourly compensation rate by scaling the weekly earnings of truck drivers by the ratio of hourly compensation to weekly earnings in the whole Trucking and Warehousing SIC. Specifically, assuming that the transportation and materials- moving profession ( BLS occupation data of Table 5- 11) corresponds to the
18The NIPA and the BLS disagree on two wage categories: state and local government employees, and private- household employees ( Table 5- 11). I am unable to explain this discrepancy.
38
SIC for trucking and warehousing, I estimate the full compensation rate for drivers of trucks as:
ACtd= AWEtd⋅ HCtwAWEtm eq. [ 5- 12]
where:
ACtd = the average compensation rate for drivers of light- duty or heavy- duty trucks ($/ hour)
AWEtd = the average weekly earnings of drivers of light- duty or heavy- duty trucks ($ 377/ week for drivers of LDTs, $ 477/ week for drivers of HDTs; Bureau of Labor Statistics, unpublished tabulations, 1993)
HCtw = the hourly compensation rate in the trucking and warehousing industry ($/ hour, from Table 5- 11; row: Trucking and Warehousing; column: Data from the National Income Product Accounts ( NIPA) of the United States, 1990, $/ hour compensation)
AWEtd = the average weekly earnings of all persons in the transportation and materials- moving profession ($/ week, from Table 5- 11; row: Trucking and Warehousing; column: BLS occupation data, $/ week, earnings)
Note that the average weekly wage in the trucking and warehousing industry is virtually the same as the average weekly earnings in the transportation and material moving occupation ( Table 5- 11). This gives me confidence that the NIPA estimate of total compensation in the trucking and warehousing industry is the appropriate measure of the cost of travel time in the trucking industry.
Note that, because truck drivers are paid to produce driving, they are compensated for all of the personal resources, including attention, that they must devote to driving, and hence are compensated for the pure utility or disutility of the driving experience -- a type of the hedonic cost mentioned above. If driving were much more demanding and stressful than it actually is, drivers would be paid more; if it were virtually effortless, they would be paid much less. ( By contrast, the value of the productivity foregone by the business traveler does not, by definition, include the disutility of the driving.). This means that the “ extra” hedonic cost of driving commercial trucks is zero.
Intercity and transit buses. To estimate the cost of paid travel time of passengers on intercity and transit buses, I assume that the ratio of the paid ( monetary) time cost to the unpaid ( non- monetary ) time cost for travel in buses equals the same ratio for travel in private LDAs used for business purposes ( data in Table 4- 1; non- monetary time costs are discussed in Report # 4). To estimate the cost of the bus driver’s time, I use equation 5- 12 but with the variable AWEtd defined to be the average weekly earnings of bus drivers in 1990, as reported by the Bureau of Labor Statistics ($ 394/ week; BLS, unpublished tabulations, 1993).
39
Police vehicles. I assume that the value of police activities foregone on account of travel in police vehicles is the full hourly compensation rate for police officers. I estimate the full compensation rate for police officers as I estimate it for truck drivers:
ACp= AWEp⋅ HCaAWEa eq. [ 5- 13]
where:
ACp = the average compensation rate for policeman ($/ hour)
AWEp = the average weekly earnings of police and detectives ($ 553/ week; Bureau of Labor Statistics, unpublished tabulations from the Current Population Survey, 1993)
HCa = the average full hourly compensation rate of all employees ($/ hour, from Table 5- 11; row: all employees; column: Data from the National Income Product Accounts ( NIPA) of the United States, 1990, $/ hour compensation)
AWEa = the average weekly earnings of all workers ($/ week, from Table 5- 11; row: all employees; column: BLS occupation data, $/ week, earnings)
Deduction to avoid double counting the cost of police- officer time in patrol cars. In Report # 7, on government expenditures on motor- vehicle goods and services, I have estimated the cost of all police time -- including time in police cars -- devoted to patrolling highways, enforcing traffic laws, and preventing and investigating motor- vehicle related crimes. The cost of police travel time that is part of the total motor- vehicle police cost of Report # 7 double counts some of the cost of police travel time estimated here. But how much is double counted?
Here, we estimate the cost of all police time in motor vehicles. In Report # 7, our estimates of police costs related to motor- vehicle use include, implicitly, the cost of police travel time that is related one way or another to the public’s use of motor vehicles. Thus, the question becomes: what fraction of total police travel time is related in anyway to the public’s use of motor vehicles -- for that fraction already is included in the estimates of Report # 7, and hence should be deducted here. In Report # 7, I estimate that about 30% of the total expenditures on police ( all police activities, for all purposes) can be attributed to the public’s use of motor vehicles. On the basis of this, I assume that 30% of the total cost of police time in police vehicles already is counted in my estimate of police expenditures attributable to the public’s use of motor vehicles in Report # 7 ( summarized in Table 1- 7 of Report # 1). Thus, I here deduct 30% of total police time cost in travel.
I recognize, but do not analyze, the possibility of double counting ( once in Table 1- 7, and once again in Table 1- 9 or 1- 4, of Report # 1) the time spent in fire vehicles and other public vehicles ( e. g., public cars driven by judges) used for motor- vehicle related purposes, such as putting out car fires, or trying cases involving drunken driving.
40
I also assume, for the reasons discussed above in relation to truck- driver time, that the hedonic cost of time in police cars is zero.
5.9 OVERHEAD COSTS OF BUSINESS, TRUCKING, AND GOVERNMENT FLEETS
Fleets have several kinds of “ overhead” costs on top of the costs analyzed in the preceding sections. For example, the total operating costs in SIC 421, “ Trucking and Courier Services,” include lease and rental of buildings and non- transportation equipment, fuel for heat and power, salaries of management and office staff, insurance for and maintenance of buildings and nontransportation equipment, and drug and alcohol testing programs ( Bureau of the Census, Motor Freight Transportation and Warehousing Survey: 1991, 1993). Large Federally owned fleets also have similar overhead costs ( Frisbee, 1994). As shown in Table 5- 3, these overhead costs are a substantial fraction of total operating costs.
In Table 5- 3 , the difference between the “ net” operating cost, which includes only direct transportation costs ( vehicles, fuel, drivers, insurance, maintenance and repair... no overhead), and total cost including overhead ( buildings, equipment, electricity..) is $ 0.20 to$ 0.25 per mile. I assume that all of this difference ($ 0.05/ mile) is a cost of the motor- vehicle fleet, and then multiply it by total VMT by fleet vehicles ( calculated as:, from equation 5- 9b) to obtain an estimate of the total dollar overhead cost of fleets. I also assume that the cost per mile includes any interest charges pertinent to any long- lived capital. VMTv⋅ FVMTvvΣ
Whether or not a particular overhead cost should be counted as a cost of motor- vehicle use depends on whether or not the cost would be different, by some measure, if a different transportation mode were used. For example, one can argue that any freight- shipping concern, regardless of the mode of shipment that it employs, requires buildings and office supplies and accountants, and hence that the cost of these should be attributed to freight movement in general, not to any particular mode of shipment. I believe, though, that the exact amount of this overhead ( measured in dollars per ton or ton- mile shipped, dollars per dollar of revenue, or something similar) probably does vary, if only slightly, from mode to mode, and so technically is a cost of each particular mode. I have included overhead costs in this analysis. ( Note that overhead does not include in- house maintenance and repair at business and government fleets; this is counted separately above under “ parts, supplies, maintenance...” It also does not include the administrative cost of self- insurance for motor vehicles, which again is counted separately elsewhere in this report.)
5.10 PRIVATE MONETARY COSTS OF MOTOR- VEHICLE ACCIDENTS
41
5.10.1 Background
In 1991, motor vehicle accidents damaged nearl