The publications in this collection do
not reflect current scientific knowledge
or recommendations. These texts
represent the historic publishing
record of the Institute for Food and
Agricultural Sciences and should be
used only to trace the historic work of
the Institute and its staff. Current IFAS
research may be found on the
Electronic Data Information Source
site maintained by the Florida
Cooperative Extension Service.
Copyright 2005, Board of Trustees, University
January 1983 Bulletin 833
Economic Impact of Empty Backhauls in
Florida Fresh Fruit and Vegetable
Richard L. Kilmer, Henry E. Ramirez, and
Forrest E. Stegelin
f.I AY 2 1983
I.F.A.S. Univ. of Florida
Agricultural Experiment Stations
Institute of Food and Agricultural Sciences
University of Florida, Gainesville
F. A. Wood, Dean for Research
INTRODUCTION .................................... 1
OBJECTIVES ................................... 1
DETERMINATION OF AVERAGE EMPTY BACKHAUL
MILES PER TRUCKLOAD ................................ 2
Sampling Technique ...................................... 2
Sample Characteristics .................................... 4
Determination of Total Empty Backhaul Miles,
April 1979 ................................ .......... 7
POTENTIAL FOR LOADED BACKHAULS .................... 8
ECONOMIC IMPACT OF EMPTY BACKHAUL MILES ........... 8
SUMMARY AND CONCLUSIONS .......................... 10
APPENDIX A .......................................... 11
Factors Affecting Backhaul ................................ 11
Model Estimation ........................................ 13
REFERENCES ......................................... 16
Richard L. Kilmer and Forrest Stegelin are assistant professors of Food
and Resource Economics. Henry E. Ramirez is a former graduate student in
Food and Resource Economics.
Fuel and energy conservation concerns have placed new emphasis on the
problem of empty backhauls and their impact on the operating efficiency and
productivity of the motor carrier transport industry [Moser, 1978; Williams,
1977]. Empty backhauling creates resource use inefficiencies which result in
higher costs to shippers, lower farm prices, and higher social costs through
higher fuel consumption.
In 1979, transportation costs were relatively high for fresh fruits and veg-
etables (FF&V), representing 10 to 25 percent of the retail price for selected
produce [USDA-ESCS, 1980, p. 48]. Fuel prices were the leading factor in
increasing transportation costs during 1979, accounting for 20.5 percent of
the total cost per mile in January 1979 and rising to 29 percent by January
1980 [USDA, ESCS, 1981, p. 88]. Diesel fuel prices paid by independent
truckers increased by 114 percent between June 1976 and February 1980,
with about two-thirds of that increase occurring after January 1, 1979 [USDA,
ESCS, 1981, p. 88]. Rates paid for shipping FF&V increased by 10 to 15
percent during 1979 [USDA-ESCS, 1981, p. 88].
Trucking dominates FF&V transport in Florida. In 1979, only one percent
of the perishable traffic from Florida was by rail, whereas truck transporta-
tion accounted for 99 percent of Florida FF&V shipments [Stegelin, 1981, p.
2]. Although no exact figures exist, we estimated that 30 to 50 percent of the
trucks hauling Florida FF&V experienced empty backhauls.' One method of
improving the cost efficiency of food distribution is to decrease the empty
The objective of this study was to determine the economic impact on
truckers and farmers of empty truck backhauls in the transport of Florida
FF&V Specific objectives were:
1. to determine the average empty backhaul miles per truckload in the
transportation of Florida FF&V;
2. to determine the total empty backhaul miles for all truckloads during
an average month in the transportation of Florida FF&V;
3. to determine the interstate potential for finding commodities to
backhaul to Florida; and
4. to determine the potential reduction in transportation costs and the
potential improvement in farm prices.
1. Estimated after conversations with several Florida truck brokers and Florida
agricultural inspection officers.
DETERMINATION OF AVERAGE EMPTY
BACKHAUL MILES PER TRUCKLOAD
In order to determine the amount of fuel consumed in empty backhauling
by Florida FF&V trucks, one must identify trucks that haul FF&V from
Florida, and then follow the return trip to observe the backhaul miles traveled
No secondary data were available on the number of empty backhaul miles
experienced by truckers hauling Florida FF&V. A backhaul questionnaire
was developed and distributed during April 1979 by the Transportation Research
Center, Department of Civil Engineering, University of Florida, with the
assistance of Henry Ramirez of the Food and Resource Economics Depart-
ment, Institute of Food and Agricultural Sciences. The questionnaire [Pav-
lovic, 1980, pp. A14-16] was designed to collect information from each trucker
on: (1) the unload destination of Florida FF&V, (2) the distance traveled
loaded or partially loaded during the backhaul, (3) the percent utilization of
capacity during a partially loaded portion of the backhaul, and (4) the dis-
tance traveled empty during the backhaul.
Sampling was performed in April 1979 in order to distribute the question-
naire before the peak and slack months for Florida FF&V shipments. April
was a representative month for those months with volumes greater than the
average (Figure 1). It was assumed that April was an average month for
manufactured product flows into Florida which could be used as backhauls.
The survey was conducted at Florida Agricultural Inspection Stations on
three interstate highways: I-10, 1-75, and 1-95. These stations accounted for
approximately 81 percent of the total fresh citrus traffic passing all stations
(16 total stations) during the 1978-79 marketing season and 83 percent during
April 19792 [Florida Department of Agriculture and Consumer Service, 1978-
79]. Sampling took place on the southbound (1-75, 1-95) and eastbound
(I-10) lanes of traffic to assure sampling of truckers returning to Florida after
hauling a load of FF&V out of the state. By sampling the southbound/east-
bound lanes it was possible to ask the trucker if he/she had just hauled a load
of FF&V out of Florida on his/her last trip. If so, the trucker was included
in the sample.
Because of the amount of traffic, safety hazards, and the truckers' desire
for quick inspection, it was not possible to randomly select carriers. There
was no assurance of getting a questionnaire to a carrier once it had been
selected. Also, only with the inspector's help could all types of trucks that
2. No similar figures exist through the Agricultural Inspection Stations on the
distribution of fresh vegetables.
1976-77 to 1979-80 Four-Year Average --------
S 500 1976-77 to 1979-80 Monthly Average
200 -- 1978-79 Marketing Season
I II III II II
Sept. Oct. Nov. Dec. Jan. Feb. March April May June July Aug.
Figure 1. Monthly Florida fresh fruit and vegetable shipments, 1976-77 to 1979-80. Source: Federal-State Market News Service,
haul FF&V have been sampled. The inspectors were very knowledgeable
about the different carriers, but they were not always available to provide
assistance. Therefore, sampling efforts were concentrated on stopping each
refrigerated van that passed. These vans are used almost exclusively except
during seasonal periods when truck demand is extremely high [Pavlovic, et
al., 1980, p. 59]. Nonrefrigerated vans are sometimes used for shorter dis-
tance markets, and the commodities hauled are iced. Open vans are also used
but usually only in special instances, such as for hauling watermelons.
Four out of five times, sampling took place on Sunday night, anywhere
from 3 p.m. to 3 a.m. (Table 1). On Easter Sunday, truck traffic was light
and few questionnaires were passed out. Therefore, sampling was also done
on that Monday night.
The sample obtained (55 returned questionnaires of which 48 were use-
able) was small compared to the total number of Florida FF&V truckloads
(Table 1)(17, 373)3 going interstate during April 1979 to domestic destina-
tions east of the Mississippi (not including Miami) plus Toronto and Montreal,
Canada. While the sample was small, we can address the question of credi-
bility of its use.
In 1978-79, 86.8 percent of Florida's FF&V unloads were taken to desti-
nations east of the Mississippi River (including Miami) plus Toronto and
Montreal, Canada [Pavlovic, et al., 1980, pp. 34-35]. In April 1979, 85.98
percent of the unloads were east of the Mississippi River (including Miami)
plus Toronto and Montreal, Canada [USDA Agricultural Marketing Service].
Thus, the results of this sample were representative of trucks hauling Florida
FF&V east of the Mississippi River where most of Florida's FF&V were
A trucker's utilization of capacity in backhaul operations may be affected
by the length of fronthaul (one-way mileage), commodities which may be
hauled, intermediate points which may be served (implying a potential for
decreasing empty backhauls), product availability and direction of movement
into and from regions of carrier's domicile, owner-operator versus fleet and/or
leasing, and time in business (experience) [Felton, 1978; Pederson, et al.,
1980]. In our model, the only variable found to significantly affect the percentage
of empty backhaul mileage was one-way fronthaul mileage (Appendix A),
indicating that when a sample of truckers is taken, the sample should be weighted
according to the geographical distribution of Florida FF&V. This factor was
found to be a characteristic of the sample used in this study, and it is discussed
3. Calculated from Federal-State Market News Service.
Table 1. Backhaul questionnaire distribution.
Percent Questionnaires Percent fresh citrus
Distribution Date Questionnaires of received useable shipped
location distributed total (useable) of total (percent)
Interstate 95 April 22, 1979
Yulee, Florida (Sunday) 153 46.7 31 56.2 56.4
Interstate 75 April 1, 1979 and
White Springs, Florida (Sunday) 112 34.3 18 31.3 26.2
April 15, 1979 (15)
Interstate 10 April 15, 1979 and
Live Oak, Florida (Sunday) 62 19.0 6 12.5 17.4
April 16, 1979 (6)
327 100 55 100 100
a. In April 1979, these three agricultural inspection stations handled 83 percent of the fresh citrus shipped [Florida Department of Agriculture and Consumer
b. From Pavlovic, et al., 1980, p. A-4.
c. Source: Florida Department of Agriculture and Consumer Service, 1978-79. The distribution of vegetables is not available.
In order to ascertain whether the sample of questionnaires reflected the
geographic distribution of unloads among states (Table 2), two distribution-
free tests for central tendency were performed: (1) median distribution-free
signed rank test (Wilcoxon) [Hollander and Wolfe, 1973, pp. 27-33], and (2)
dispersion a distribution-free rank test (Ansari-Bradley) [Hollander and Wolfe,
1973, pp. 83-93]. It was found that the percentage distributions of questionnaires
and unloads were not significantly different at the 99 percent confidence level.
This finding indicated that even though the fronthaul destination of truckers
hauling Florida FF&V was not controlled when the questionnaires were
distributed, the percentage distribution of the questionnaires and unloads was
not statistically different. Thus, the information contained in the sample appeared
representative of the geographical distribution of the destinations to which
truckers hauled Florida FF&V in April 1979.
The sample mean was used to make projections to the population. The
precision of the sample mean of empty backhaul miles per truck when used to
Table 2. The percentage distribution of useable questionnaires and Florida FF&V
unloads by state, April 1979.
State Questionnaires Unloads
Number Percent 1000 cwt Percent
Alabama 2 4.7 84 2.2
Delaware (1)a ..
Georgia 1 2.3 239 6.1
Illinois 2 4.7 240 6.2
Indiana 1 2.3 72 1.8
Kentucky 1 2.3 42 1.1
Maryland; Washington, D.C. 2 4.7 290 7.4
Massachusetts 3 7.0 371 9.5
Michigan 1 2.3 162 4.2
North Carolina (2)a -
New York City; New Jersey 11 25.6 872 22.4
Ohio 5 11.6 244 6.3
Pennsylvania 3 7.0 515 13.2
Rhode Island 1 2.3 17 .4
South Carolina 2 4.7 128 3.3
Tennessee 0 0 74 1.9
Virginia (l)a ..
Wisconsin 3 6.9 26 .7
Montreal 1 2.3 184 4.7
New Brunswick (1)a --....
Toronto 4 9.3 337 8.6
TOTAL 43 100 3897 100
Source: USDA Agricultural Marketing Service, 1979
a. States not included in statistical tests because they have no unloads.
estimate the population mean was computed as follows. Given the sample
1. sample size (n = 48),
2. sample mean (Y = 364.9 miles),
3. the standard deviation of the sample (S = 353.3), and
4. the size of the population (N = 17,373),
there was a 95 percent probability that the population average empty backhaul
mileage per truck hauling Florida FF&V during April 1979 was between
263.1 miles and 466.7 miles.
Thus, we were 95 percent confident that the sample mean was, at most, 28
percent larger (smaller) than the population mean.
Determination of Total Empty Backhaul Miles, April 1979
The total empty backhaul mileage (EBM) for all Florida FF&V truckers
that hauled produce during April 1979 will now be estimated. In estimating
the population average, N (the number of truckloads of Florida FF&V during
April 1979) was equal to 17,373 trucks.4 Y (the average number of empty
backhaul miles per truck) was 364.9 miles,5 S (standard deviation of the sample)
was 353.3 miles,5 and n (the number of trucks in the April 1979 sample) was
48.5 The total empty backhaul mileage for Florida FF&V truckers during April
1979 was determined as follows:
EBM = NY
= (17,373) x (364.9) (I)
= 6,339,408 miles
Given that the 6,339,408 miles was a sample estimate, we were 95 percent
confident that the actual number of empty backhaul miles was between 4,570,004
and 8,108,812 miles6 for an error of estimation of 28 percent. The empty
mileage represented 35.3 percent of the total fronthaul mileage for April 1979.7
Determination of the empty backhaul miles per truck (364.9 miles) and the
total empty backhaul miles for April 1979 (6,339,408 miles) accomplished
Objectives 1 and 2. Next, the potential for loaded backhauls to Florida was
4. Calculated from Federal-State Market News Service.
5. Calculated from backhaul survey.
6. EBM 2[(N2)(S' n)([N n]IN)]/2 [Mendenhall. 1971, p. 38].
7. The average fronthaul mileage per truck from the backhaul survey was 1033
POTENTIAL FOR LOADED BACKHAULS
The potential for a loaded backhaul from state i to Florida was measured by
comparing the exempt and nonexempt commodities exported from state i to
Florida with the commodities imported from Florida to state i.8 Twenty-one
(4) states east of the Mississippi exported (imported) more commodities to
(from) Florida than were imported (exported) from (to) Florida (Table 3).
Thus, the potential existed for FF&V truckers to find a backhaul to Florida.
Furthermore, if a trucker made a delivery to one of the four states that provided
no opportunity for a backhaul, the trucker could go to one of the adjacent
ECONOMIC IMPACT OF EMPTY BACKHAUL MILES
Empty backhaul mileage cost truckers $6,339,4089 in April 1979. The
likelihood that empty backhaul mileage can be reduced to zero is improbable,
and possibly economically inefficient; however, with deregulation and more
information on the location of backhauls, empty backhaul mileage can be
The price incidence on producer prices from a change in the marketing
margin caused by a decrease in empty backhaul cost is summarized in the
following equation [Fisher, p. 261]:
Ip = 1 (2)
1 + ep/an
where: Ip = the proportion of the marketing cost change borne by the producer
ep = the price elasticity of producer supply
a = PpIP,
Pp = producer (farm) price
P, = consumer (retail) price
n = the price elasticity of consumer demand.
The price incidence on consumer prices was defined as:
Ir = 1 -Ip (3)
where: Ir = the proportion of the marketing cost change borne by the consumer.
Assuming that the price elasticity of supply was zero (inelastic supply of
fruits and vegetables in the market),"' then the proportion of price incidence
8. The Census of Transportation for nonexempt commodities is only taken during
the years ending in 2 and 7. Thus. 1972 data were used.
9. Calculated as follows: 17,373 trucks times 364.9 average empty backhaul
miles per truck times $1.00 per mile [Boles. 1980, p. 10] equals $6,339,408.
10. A viable assumption in the short run due to the short storage life for fresh
fruits and vegetables and the extended period of time required to change farm level
production plans which cause changes in market supply.
Table 3. Net exports of exempt and nonexempt commodities from individual states
to Florida, April 1972.
States Net exports
New Hampshire (-1,097)b
New Jersey 72,964
New York 25,123
North Carolina 120,778
Rhode Island 5,051
South Carolina (-2,031)b
West Virginia 25,143
Net exports 1,121,266
Note: The Census of Transportation for nonexempt commodities is only taken during the
years ending in 2 and 7. Thus, 1972 was used for.both exempt and nonexempt
commodities [U.S. Bureau of Census, 1975].
a. Ramirez, p. 36
b. Ramirez. p. 34
on producers from a decrease in empty backhaul mileage was 1 (one). Thus,
if there were adjustments in the market place correcting for technical inefficiencies
by decreasing empty backhaul cost, producer (farm level) prices would increase
while consumer (retail level) prices would remain unchanged.
The estimated retail value of Florida FF&V in April 1979 amounted to
$389.1 million (calculated from Federal-State Market News Service, Florida
Crop and Livestock Reporting Service, and U.S. Department of Commerce).
Transportation charges for FF&V were relatively high, representing 10 to 25
percent of the total retail value [USDA Economics, Statistics and Cooperative
Service, 1980, p. 48]. If an average of 17.5 percent was used, then April
1979 transportation charges for FF&V totaled approximately $68,092,500.
The estimated empty backhaul cost ($6,339,408) was 1.6 percent of the retail
value and 9.3 percent of the total transportation charges for Florida FF&V for
The potential savings of $364.90 per truck" from eliminating a technical
inefficiency attributed to empty backhaul miles should therefore be realized
in its entirety by the producer in terms of higher producer (farm level) prices,
given the inelastic supply of fresh fruits and vegetables at the market.
SUMMARY AND CONCLUSIONS
The purpose of this study was to determine the economic impact on truckers
and farmers of empty truck backhauls in the transportation of Florida FF&V.
To do this, it was necessary to (1) determine the average empty backhaul miles
per truck in the transportation of Florida FF&V, (2) determine the total empty
backhaul miles for all truck shipments during an average month in the
transportation of Florida FF&V, (3) determine the interstate potential for backhauls
to Florida, and (4) determine the potential reduction in transportation costs and
the potential improvement in farm prices.
Empty backhaul miles per trip represented 35 percent of the average fronthaul
mileage of Florida FF&V truckers. This amount represented 9.3 percent
($6,339,408) of the average monthly FF&V transportation bill, or $364.90
per truck trip. A reduction of empty backhaul miles to zero is virtually impossible
due to imperfect knowledge of backhauls, the cost of finding backhauls, and
the opportunity cost of using specialized equipment to haul general freight.
However, reducing the empty backhaul mileage by 50 percent could potentially
increase producer (farm level) prices $182.45 per truckload, or 0.46 cents per
pound of FF&V shipped, assuming an inelastic supply of FF&V at the market.
In conclusion, Florida's FF&V producers would benefit from reducing empty
backhaul mileage and improving technical efficiency, thus alleviating the "plight"
of the trucker. By reducing marketing costs, farm prices would be improved.
11. Calculated as follows: 364.9 average empty backhaul miles per truck times
$1.00 per mile [Boles, 1980, p. 10] equals $364.90 average cost of empty backhaul
miles per truck. The potential savings of $364.90 per truck equals 0.92 cents per
pound per truckload (39,479 pounds calculated from Federal-State Market News
Factors Affecting Backhaul
Because of limited information on the characteristics of truckers hauling
Florida FF&V, it was not possible before the sample was taken to stratify the
sample and statistically determine if certain hypothesized factors did influence
empty backhaul mileage. In order to perform a post-survey test on certain
factors, a multiple regression analysis of factors hypothesized to influence
empty backhaul mileage was performed.
"Empty"' backhaul mileage after a fronthaul of fresh fruit and vegetables
to an interstate destination has been discussed. Alternatively the problem was
to look at "full" backhaul mileage which was called backhaul capacity utilization
(BCU), defined as one minus the ratio of empty backhaul mileage divided by
the fronthaul mileage. This ratio was used because it was not the magnitude
of empty backhauls itself that mattered, but the magnitude of empty backhaul
mileage relative to the fronthaul distance that a trucker used in making decisions.
For example, trucker A, who traveled 2,100 miles in a fronthaul, may well
need to travel 300 miles before finding a backhaul load. Trucker B fronthauls
300 miles and may know of another load waiting for him at his origin. Instead
of looking or waiting for a backhaul, trucker B may decide to return empty.
Both truckers have an empty backhaul mileage of 300; however, the backhaul
capacity utilization of truckers A and B is quite different 85.7 and zero
percent. When empty backhaul mileage is compared with fronthaul mileage,
the economic cost of the empty backhaul relative to the economic cost of 100
percent empty backhaul is brought into perspective. The dependent variable
in the analysis is backhaul capacity utilization (BCU).
The selection of the hypothesized determinants influencing backhaul capacity
utilization of trucks hauling Florida FF&V was constrained by the information
contained in the backhaul survey. A trucker's backhaul capacity utilization
may be affected by the length of fronthaul (one-way mileage), commodities
which may be hauled, intermediate points which may be served (implying a
potential for decreasing empty backhauls), product availability and direction
of movement into and from regions of carrier domicile, owner-operator versus
fleet and/or leasing, and time in business (experience) [Felton, 1978; Pederson,
et al., 1980]. However, survey data were not available on all of these determinants.
A discussion of each of the variables included in the model follows.
Fronthaul (one-way mileage) was included as a factor that influences BCU
because attempts to avoid long empty movements on backhaul has an economic
basis (reduced cost) and participation in the carriage of commodities as
backhauls will likely increase as the length of haul increases [Paxson, 1979;
Pederson, et al., 1980]. Furthermore, the 1977 ICC study reporting loaded
empty truck mileage data implied the relative importance of mileage as a
backhaul consideration. Fronthaul mileage (FHM) was hypothesized to have
a positive effect on backhaul capacity utilization (Figure Al). When the
t^ Curve A
". Curve B
Figure AI. The relationship between fronthaul mileage and backhaul capacity
fronthaul was short, the dollar cost of the time spent to find, load, haul and
unload the backhaul may be greater than the total revenue gained. This factor
may influence the case for fronthauls of a few hundred miles, in which BCU
is almost zero for the initial proportion of the fronthaul mileage.
After the first few miles, the opportunity cost of finding a backhaul becomes
less than the total revenue gained from the backhaul, causing the BCU to
increase rapidly as the fronthaul increases (Figure Al). As the fronthaul mileage
increases, the BCU begins to increase at a decreasing rate. Truckers will
likely have some empty backhaul mileage even on long trips. BCU will be
near 100 percent because of economic incentives for a loaded backhaul.
An owner-operator variable (own) was also included because studies have
indicated owner-operators were the drivers in 21 percent of the empty vehicles
[ICC, 1977]. Fleet operators may possess more information concerning the
location of backhauls because the business can hire personnel to locate it. On
the other hand, an owner-operator spends most of his time driving the truck
and has less time to gather information of value to his business, i.e., the
location of backhauls. The hypothesized effect of an owner-operator on BCU
was negative (Figure Al, curve B). Fleet operators had a positive effect on
A ratio depicting product availability and direction of movement by truck
into and out of Florida was calculated. The ratio (PAV) was defined as the
sum of the exempt and nonexempt (or regulated) tonnage hauled from another
state to Florida divided by the sum of the exempt and nonexempt tonnage
hauled or exported from Florida to another state. A ratio greater than 1.0
inferred more carriage by volume was being shipped into Florida, thereby
implying an increased potential for obtaining a backhaul into Florida. The
hypothesized effect on BCU was positive (Figure Al, curve C).
Interstate highway corridors serving Florida FF&V truckers were included.
Two prominent highway arteries serving Florida were the interstate systems
of 1-95 connecting the Florida east coast with the Northeast and 1-75 connecting
Florida with the Midwest and the upper Midwest (via 1-65 as a feeder artery).
The relative importance of each corridor has been shown earlier in terms of
outbound shipments and citrus inspections (Table 1), but the hypothesized
effect was unknown. The hypothesized model was summarized as follows:
BCU = f(FHM, OWN, PAY 1-95) (Al)
where: BCU = backhaul capacity utilization,
FHM = fronthaul mileage,
PAV = product availability for a backhaul,
1-95 = an interstate corridor used by Florida truckers,
OWN = owner-operator, truck ownership.
The model specification incorporated a functional form that was similar to
Figure Al. It was assumed that the error term of the model was heteroskedastic
and must be corrected. The following model was specified:
CURi = Po + p, FHMi + 2, PAVi + P, 1-95 (A2)
+ P4 OWNi + ei
where: CURi = In ((100/BCUi) 1)',
In = natural logarithm,
100 = 100 percent, the upper limit of BCU,
BCUi = (1 (empty backhaul mileage/one-way fronthaul mileage))
x 100, or the percent of BCU for truck i,
FHMi = fronthaul mileage from Sanford, Florida to the interstate
destination of truck i,
PAVi = the exempt truckloads plus the nonexempt truckloads capable
of being hauled in refrigerated vans (Ramirez, p. 58-60)
from an interstate destination to Florida, divided by the exempt
truckloads plus the nonexempt truckloads capable of being
hauled in a refrigerated van from Florida to the interstate
destination of truck i,
1. When BCUi = 0, CURi is undefined. Thus, when BCUi = 0, the data entered
for BCUi = .001.
1-95i = a binary variable that equalled one for truck i if it traveled
1-95; zero otherwise,
1-75i = a binary variable that equalled zero for truck i if it traveled
on 1-65 and/or 75 (deleted for estimation of binary variables),
OWNi = a binary variable that equalled one for an owner-operated
truck i; zero otherwise,
Other = a binary variable that represented truck i if it was not owner-
operated, and was always zero (deleted for estimation of
P0, ... 4 = estimated parameters,
ei = ith error term.
The logistic function was used because it was thought most appropriate to
depict the level of backhaul at various fronthaul mileages, asymptotically
approaching both zero backhaul utilization and complete or 100 percent
backhauls. As a trucker fronthauls to destinations within close proximity to
the fronthaul origin, it was hypothesized that the economic need for a backhaul
was not as important as a backhaul was to a fronthaul of several hundred
It was assumed that the expected value2 of each disturbance term was:
E(|t) = .2 f(FHM/) (A3)
In order to correct for heteroskedasticity, the following model was estimated:
1 FHM, PAVi
(CURii/i) = fo( +i) + al+i + PA2 i +
1-95i OWN, ei (A4)
a +04 "+
where: all terms were as previously defined except ei which was determined
by estimating equation A2. Graphical observation of the errors from model
A2 indicated that errors were much larger in the medium fronthaul distances
and were relatively small in the short and long fronthaul distances. The resulting
parameter estimates and standard errors of the parameters from equation A4
were put back into equation A2 which resulted in:
CUR = 7.4557 .0069 FHM + .1388PAV +
(2.6706) (.0018) (.2153)
.0317 1-95 .6691 OWN (A5)
2. The implicit function of the error covariance matrix was related to one-way
fronthaul mileage and was assumed to be as follows: Ii l = fFHMi + ,2 (FHMi)2
+ Ii. The heteroskedasticity problem was corrected by dividing the original equation
by the absolute value of ei.
Thus, fronthaul mileage (FHM) was the only variable to enter the model
significantly.3 We concluded that the fronthaul mileage was important to Florida
FF&V truckers. A trucker will increase the percentage of loaded backhaul
miles as the one-way mileage from Florida increases.
3. The coefficient should be at least two times its standard error when a generalized
least squares estimator is used.
Boles, Patrick P 1980. Owner-Operator Costs of Hauling Fresh Fruits and Vegetables
in Refrigerated Trucks. USDA Economics, Statistics and Cooperative Service
Federal-State Marketing News Service. Florida Shipments of Fresh Fruits and
Vegetables. Winter Park, Selected issues.
Felton, John Richard. 1978. The Impact of Rate Regulation Upon Common Carrier
Backhauls. Department of Agricultural Economics Report 88, (September).
Lincoln: Univ. of Nebraska.
Fisher, B. S. 1981. A Change in Marketing Charges: Who Loses What? Am. J.
Agric. Econ., 63 (2): 261-3.
Florida Crop and Livestock Reporting Service. 1978. Florida Agricultural Statistics:
Citrus Summary, December 1978. Orlando, Florida.
Florida Department of Agriculture and Consumer Service, Division of FF&V
Inspection. Season Annual Report. Winter Haven, Selected issues.
Hollander, Myles, and Douglas A. Wolfe. 1973. Nonparametric Statistical Methods.
New York: John Wiley and Sons.
Interstate Commerce Commission. 1977. Empty/Loaded Truck Miles on Interstate
Highways During 1976. Bureau of Economics and Bureau of Operations (April),
Mendenhall, William, 0. H. Lyman, and Richard L. Scheaffer. 1971.
Elementary Survey Sampling. Belmonth, California: Duxbury Press.
Moser, Davis E. 1978. Regulatory Policy Issues in Truck Transportation.
Transportation Task Force.
Pavlovic, Karl R., Gary Long, Deborah P. Reaves, and Thomas H. Maze. 1980.
Domestic Transportation for Florida Perishable Produce. Dept. Civ. Eng.
Transport. Res. Cen. (November). Gainesville: Univ. of Florida.
Paxson, D. S. 1979. Motor Carrier Deregulation and the Opportunities for Reducing
Empty Truck Mileage, AAR Truck and Waterway Information Center Staff
Paper 79-2 (June 4). Wash., D.C.
Pederson, Larry E., R. C. Mittelhammer, and Kenneth L. Casavant. 1980. Factors
Affecting Interstate Backhauling of Exempt Agricultural Commodities by
Regulated Motor Carriers: A First Look. Transportation Journal (Summer) pp.
Ramirez, Henry Eugene. The Impact of Empty Backhauls on the Fuel Consumed
in Truck Transportation of Florida Fresh Fruits and Vegetables. Masters thesis,
Univ. of Florida, 1980.
Stegelin, Forrest E. 1981. Agricultural Transportation Issues for Florida Perishables
in the 1980s. Florida Food and Resource Economics 39 (March-April).
Gainesville: Univ. of Florida.
U.S. Bureau of the Census, Transportation Division. 1975. 1972 Census of
Transportation, Commodity Transportation Survey, Public Use Tape Users
Manual. Washington, D.C. (August).
U.S. Department of Commerce. Labor Statistics. Washington, D.C., Selected
U.S. Department of Agriculture Agricultural Marketing Service, Fruit and
Vegetable Division. Fresh Fruit and Vegetable Unloads. Washington, D.C.,
U.S. Department of Agriculture Economics, Statistics and Cooperative Service.
1980. Developments in Marketing Spread for Food Products in 1979. Agric.
Econ. Rep. 449, Washington, D.C.
U.S. Department of Agriculture Economics, Statistics and Cooperative Service.
1981. Agricultural-Food Policy Review: Perspective for the 1980s. Agricultural
Food Policy Report 4, Washington, D.C.
Williams, Andy. 1977. Agricultural Transportation: Highway Blues.
This public document was promulgated at an annual cost of $1980
or a cost of 990 per copy to provide information on the economic
impact on truckers and farmers of empty truck backhauls in the
transport of Florida fresh fruits and vegetables.
All programs and related activities sponsored or assisted by the Florida Agricultural
Experiment Stations are open to all persons regardless of race, color, national origin,
age, sex, or handicap.