Front Cover
 Title Page
 Table of Contents
 List of Tables
 Conceptual model
 Data collection and truck market...
 Empirical model
 Findings from the full/empty truck...
 Summary and conclusions
 Back Cover

Group Title: Bulletin (Technical) - Agricultural Experiment Stations, University of Florida - 862
Title: Full-empty truck movements into Florida
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00027300/00001
 Material Information
Title: Full-empty truck movements into Florida
Series Title: Bulletin Agricultural Experiment Stations, University of Florida
Physical Description: 37 p. : ill. ; 23 cm.
Language: English
Creator: Beilock, Richard
Kilmer, Richard L
Publisher: Agricultural Experiment Stations, Institute of Food and Agricultural Sciences, University of Florida
Place of Publication: Gainesville Fla.
Publication Date: 1986
Subject: Trucking -- Florida   ( lcsh )
Freight and freightage -- Florida   ( lcsh )
Farm produce -- Transportation -- Economic aspects -- Florida   ( lcsh )
Genre: government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
non-fiction   ( marcgt )
Bibliography: Bibliography: p. 35-37.
Statement of Responsibility: Richard Beilock and Richard L. Kilmer.
General Note: "December 1986."
Funding: Bulletin (University of Florida. Agricultural Experiment Station) ;
 Record Information
Bibliographic ID: UF00027300
Volume ID: VID00001
Source Institution: Marston Science Library, George A. Smathers Libraries, University of Florida
Holding Location: Florida Agricultural Experiment Station, Florida Cooperative Extension Service, Florida Department of Agriculture and Consumer Services, and the Engineering and Industrial Experiment Station; Institute for Food and Agricultural Services (IFAS), University of Florida
Rights Management: All rights reserved, Board of Trustees of the University of Florida
Resource Identifier: aleph - 000952422
oclc - 17257207
notis - AER4743

Table of Contents
    Front Cover
        Front Cover
    Title Page
        Page i
        Page ii
    Table of Contents
        Page iii
    List of Tables
        List of Tables
        Page 1
    Conceptual model
        Page 2
        Single committed carrier
            Page 3
            Page 4
            Page 5
            Page 6
            Page 7
            Page 8
            Page 9
            Page 10
        Multiple committed carriers
            Page 11
            Page 12
    Data collection and truck market overview
        Page 13
        Data collection
            Page 13
            Page 14
        Truck market overview
            Page 15
            Page 16
    Empirical model
        Page 17
        Market conditions
            Page 18
            Page 19
        Carrier characteristics
            Page 20
            Page 21
            Page 22
    Findings from the full/empty truck model
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
    Summary and conclusions
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
    Back Cover
        Page 38
Full Text


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
of Florida

l |

Full-Empty Truck Movements

Richard Beilock and Richard L. Kilmer

Central Science

OCT 231987

University of Florida

Agricultural Experiment Stations
Institute of Food and Agricultural Sciences
University of Florida, Gainesville
J. M. Davidson, Dean for Research

Bulletin 862 (technical)

,er 1986




Richard Beilock and Richard L. Kilmer

The Authors

Richard Beilock and Richard L. Kilmer are Assistant and
Associate Professors in the Food and Resource Economics Depart-
ment at the University of Florida.



LIST OF TABLES .............................................iv

LIST OF FIGURES .............................................iv


CONCEPTUAL MODEL .............................................2
Single Committed Carrier ................................ 3
Multiple Committed Carriers............................11

Data Collection.......................................13
Truck Market Overview...................................15

EMPIRICAL MODEL .............................................17
Market Conditions.....................................18
Carrier Characteristics...............................20


SUMMARY AND CONCLUSIONS .....................................29

NOTES............................. ..........................31

REFERENCES ..................................................35


Table Page

1. Percent empty vehicles by length of previous haul........9

2. The definition of variables included in Xn.............. 19

3. The results of logit analysis of full/empty southward +
movements of motor carriers of Florida produce and
ornamentals, 1982/1983 ..................................24


Figure Page

1. Representation of demands and supplies for
inbound trucking services for areas which are
net importers and net exporters of freight...............6

2. Relationship of rates, cost, net revenue
and distance............................................10

3. Monthly Florida fresh fruit and vegetable
shipments, average 1976/1977 to 1979/1980............... 14

4. Probability of full movement by month and
distance....................... ........................25

5. Probability of full movement by carrier type,
ICC authority and distance..............................27



Transportation plays a crucial role in the determination
of economic activity over time and space. Despite this fact,
relatively little attention has been devoted to understanding
transportation markets. Indeed, it is common in interregional
trade modeling to treat transportation supply as perfectly
Elastic. This tangential treatment of transportation markets
has been identified as a major reason for the failure of many
interregional trade models to reflect adequately the economic
activity being depicted (Johnson, p. 63).

One factor in the determination of price in transportation
markets is the intensity of equipment (i.e. vehicle) utiliza-
tion (Wilson, Klindworth and Brooks, Jara-Diaz, and Baesmann and
Daughety). Equipment utilization may be viewed as having two
dimensions: frequency of use (i.e. number of trips per unit
time) and ratio of full-to-empty mileage. In this bulletin the
focus is on the latter. In particular, Florida (inbound) move-
ments of motor carriers that haul produce and ornamentals (per-
ishables) from Florida are examined. Revenues generated from
inbound movements contribute to covering the costs of the round
trip. If carriers consistently must travel empty into Florida,
then outbound produce or ornamentals must bear the full cost of
the round trip.1 Kilmer, et al. (p. 10), estimated that empty
movements into Florida raise the costs of transporting Florida
perishables by almost ten percent.

Empty movements are not indicative of an inefficient sys-
tem. Interregional freight imbalances may make some empty
movements unavoidable. Alternatively, the demands for carriage
at one location may be so great as to justify rushing vehicles
to it from remote points without the delays inherent in secur-
ing and handling loads. Such empty movements are of little
concern. Rather, instances in which empty movements occur due
to some market imperfection, such as incomplete information or
regulatory constraints, are of interest. These are the empty

movements which may be reduced by alternative transportation

The purpose of this bulletin is to identify the determi-
nants of full/ empty southbound movements into Florida by those
truckers that subsequently haul Florida fresh produce and orna-
mentals out of this state. Specific objectives are (a) to
develop a conceptual framework of the decision making process
used by truckers in making a full/empty movement decision, and
(b) to develop a logit model which will quantify the determi-
nants of full/empty movements into Florida.

In the next section, a conceptual framework for depicting
carrier full/empty movement decision making is described.
Next, the data and a brief overview of the trucking market is
provided. Building from the conceptual model and the specific
market situation, an empirical model is elaborated upon and
then estimated. Finally, the results are presented and conclu-
sions drawn.


In this section, two types of operating conditions for
carriers are discussed--committed and uncommitted carriers. A
committed carrier is one which intends to travel from its cur-
rent location, i, to another location, j, regardless of whether
a load from i to j can be secured. Reasons for such behavior
include a prior commitment to pick up a load at j, the expecta-
tion of lucrative loads being available at j, and some other
obligation to go to j, such as the need to return to the car-
rier's home base for maintenance or rest. An uncommitted car-
rier is one which does not have a compelling reason to proceed
from i to j. Uncommitted carriers would require, therefore, a
load in that direction to make the movement.

In the empirical application of the model to be developed,
primary focus will be on committed carriers. Again, committed
carriers are committed to the i-to-j trip because they have

loads or the expectation of acquiring loads at j. The origin
point, i, will vary from carrier to carrier, ranging, in gener-
al, from southern Georgia to Quebec City and the Maritimes.
The destination point, j, will always be Florida.

Single Committed Carrier

SConsider a carrier committed to making a movement from
point i to point j. The carrier has, in effect, two options
for the i-to-j trip; either the carrier will take a load or he
will not. Assuming that the carrier is a profit maximizer, a
load will be sought if the returns exceed the additional costs
S associated with acquiring, handling, and transporting a load.2
It should be repeated that as the carrier is already committed
to the i-to-j movement, these additional or variable costs will
not include the full cost of the i-to-j movement. For example,
if the vehicle realizes 5.0 miles per gallon when empty and 4.5
miles per gallon when loaded, the additional or variable fuel
costs are those associated with the decline from 5.0 to 4.5
miles per gallon and not the entire fuel bill. The remainder
of the fuel bill is, in effect, a fixed cost.

The additional costs associated with an i-to-j load may be
divided into three categories:

1. search-related costs (DSC),

2. handling and transporting costs, PLU and (FRC. -
ERCij), and

3. the impact on future earnings of the delay in-
volved in acquiring and handling an i-to-j load,
(RATEij) (DPR).

Subtracting these costs from the freight rate for the i-to-j
movement (RATEij) yields the net revenue (Pij) from an i-to-j

(1) P ij RATE1j DSC PLU (FRCij ERC ) + (RATEjk)

where: RATEij = freight rate (total revenue) from i-
DSC = direct search cost
PLU = cargo pickup and drop costs
FRCij = full running (i.e., distance-
related) cost
ERCij = empty running cost
RATEjk = current freight rate (total
revenue) from j-to-k
DPRij = proportion change in RATEjk during
the delay in arrival at j from i
resulting from an i-to-j load

The carrier will not seek out an i-to-j load unless Pij is

The second righthand term, DSC, is the cost associated
with searching for a load. There is no ij subscript as these
costs are here assumed to be invariant across routings. This
assumption is made largely for expositional convenience. Re-
laxing this assumption would have no significant impact on the
model's implications. PLU is the cost of picking up and un-
loading the cargo. It includes any repositioning costs to move
from the current location to the loading dock. PLU is simi-
larly assumed to be invariant to routings. The fourth right-
hand term is the differential between full and empty running or
distance-related costs. As the time required to span distance
is essentially the same whether the vehicle is empty or full,
this cost element is comprised primarily of additional fuel and
maintenance expenses resulting from the additional weight.

For carriers not committed to the i-to-j movement, the
avoidable distance-related cost is the entire cost of moving a
vehicle and load across distance, FRCij. FRCij is likely to be
much higher than the avoidable distance-related costs for com-
mitted carriers, (FRCij ERCij). For example, Boles (1980)
estimated that the miles per gallon for a tractor-trailer drops
from 5.0 when empty to 4.5 when full. Assuming $1.20 per gal-
lon fuel cost for a 1,000 mile trip, the avoidable fuel costs
for an uncommitted and a committed carrier, respectively, are
$266.67 and $26.67 (or 26.67 and 2.67 cents per mile, respec-
tively). If rates tend toward the costs of the marginal carri-
ers (i.e., the carriers that supply the transportation services
for the residual units to be transported) and if the marginal
carriers are uncommitted carriers, then the rate-distance gra-
dient would be expected to be steeper than if the market could
be served entirely by committed carriers. The higher rate-
distance gradients would be expected to occur for shipments
into areas such as Florida that are net importers of freight.

This point is illustrated in Figure 1. The market supply
curve for inbound trucking services, S is formed by arraying
the lower cost, committed carriers in the OA segment, and the
higher cost, uncommitted carriers beyond A. It is assumed that
rates correspond to the costs of the marginal carriers. If
region j is a net exporter of freight, then at least some vehi-
cles must travel toward region j empty. In such a case, the
marginal carrier for freight inbound to region j is a committed
carrier (i.e., a carrier committed to travelling to region
j). The demand for inbound trucking services, DE, therefore,
will intersect the supply curve to the left of A. If on the
other hand, the region is a net importer of freight, the mar-
ginal carrier is the uncommitted carrier; hence the demand for
inbound trucking services, DI, would intersect the supply curve
to the right of A.

It should be stressed that the discussion of Figure 1 is
not meant to imply that the absolute level of demand for in-
bound transport services is necessarily higher for net import-

ST = Supply of in-
bound trucking

S= Demand for inbound
trucking services
in importing regions

Quantity of trucking

Figure 1. Representation of demands and supplies for
inbound trucking service for areas which are
net importers and net exporters of freight.


ing than for net exporting regions. Rather, the marginal car-
riers for net importing regions will be uncommitted carriers,
while the marginal carriers for net exporting regions will be
committed carriers.

RATEjk is the current freight rate (i.e., the total reve-
nue) for the load that the carrier anticipates to take from
j. DPRij is the proportion change in RATEjk during the delay
resulting from having an i-to-j load. That is, if the carrier
anticipates 80 percent of the current j-to-k rate to be the
going rate upon arrival at j from i without taking an i-to-j
load, and 75 percent of RATEjk to be the going rate upon arri-

val at j with an i-to-j load, then DPRij equals -.05 (.75 .80
= .05). (RATEjk) (DPR), therefore, is the revenue foregone
(or gained if DPR were positive) on the j-to-k load due to the
time involved to acquire and handle an i-to-j load. Strictly
speaking, the carrier would look at the impact of the i-to-j
load on the entire stream of future earnings. It seems likely
that in most instances expectations would not be clearly formed
regarding the effect of taking an i-to-j load on the revenues
of all but the immediately subsequent load. Taking account of
later loads would simply involve the inclusion of more terms
similar to (RATEjk) (DPR).

A major implication of the preceding discussion concerns
the relationship between distance and the probability of a
committed carrier hauling a load from i to j. Committed and
uncommitted carriers would be expected to have comparable
distance-unrelated costs, DSC and PLU. Likewise, there is no a
priori reason to expect that (RATEjk) (DPR) would differ (also
a distance-unrelated cost), at least systematically, between
committed and uncommitted carriers. However as has already
been argued, the distance-related costs (i.e., FRC for uncom-
mitted and (FRC ERC) for committed carriers) would differ
considerably between committed and uncommitted carriers. If
rates tend toward the costs of the marginal uncommitted carri-
ers for inbound freight to net importing regions, then for

committed carriers the rate-distance gradient would be steeper
than the cost-distance gradient:

(2) a Rate a (FRC ERC)
a Distance 3 Distance

This implies that the net revenue associated with acquiring an
inbound load would be positively related to distance for com-
mitted carriers:5

(3) P 0
a Distance

The relationship is illustrated in Figure 2.

If net revenue is positively related to distance, then the
probability of a committed carrier seeking an inbound load
should also be positively related to distance. Indeed, there
exists some empirical evidence to support the hypothesized
length of haul-full movement relationship. Paxson (1979) ana-
lyzed data from the National Motor Transport Data Base. He
found that the percent of empty vehicles was negatively related
to the length of the previous haul (Table 1). As many trucks
operate in shuttle fashion between two or more points, the
length of the previous haul is a reasonable proxy for the
length of the current haul. For those vehicles heading in the
lighter direction of freight flows (i.e., the direction which
has less freight than the reverse direction), there should be
no distance-full movement relationship. However, for those
moving in the direction of the heavier flow of freight, the
relationship should be positive. Therefore, when viewing a
mixture of heavy and light freight direction movements, a posi-
tive relationship, such as that found by Paxson, would be ex-

Perhaps it should be noted that the above discussion does
not imply that carriers would seek the longest distance inbound
trip to any net importing region. First, a carrier is commit-
ted to go to a specific location j because of the opportunities
anticipated at that location. As, by definition, net importing

Table 1. Percent empty vehicles by length of previous haul.

Previous length of haul (miles)
500 1000 1500 2000 All
Carrier Type <500 1000 1500 200 2500 <2500 mileage

Regular route 10 7 3 2 N/Aa N/A 9
Common carriage

Irregular route 33 18 14 10 8 5 18
Common carriage

Private carriage 33 21 16 13 6 3 23

Contract carriage 32 19 15 11 8 6 19

Exemptb carriage 40 29 20 12 10 9 22

All types 31 20 15 11 8 6 19

Trailer Type
Regular van 24 16 12 9 7 4 17

Reefer van 34 17 12 9 7 7 13

Flatbed trailer 33 22 17 15 11 6 29

Tanker trailer 44 28 31 35 24 N/A 39

Moving van 27 20 16 5 7 7 12

Special trailers 38 22 20 15 14 N/A 20

All trailers 32 19 14 11 8 7 19

aAn "N/A" denotes that the category has no observations or only a
small number (less than 20) of observations.
bIncludes Agricultural Cooperative Hauls.

Source: Paxson [1979].

loads not sought
by the average

loads sought by the
average carrier

COST b, c

0 D \e

0 ID


Figure 2: Relationship of rates, costs, net revenue, and

a. Available to total freight rate
b. The distribution at each distance has been
collapsed to its mean.
c. Costs associate with searching for and handling
of a load
d. (Expected net revenue)
e. Distribution across carriers

RATE a, b

regions have less outbound than inbound freight, a committed
carrier probably has some fairly specific reason for anticipat-
ing favorable treatment there. Second, like migrant farm work-
ers, produce/ornamentals haulers usually serve different areas
depending upon the season. Therefore, even if the carrier
wanted to maximize distance, the optimal location for a base of
operations would not be a simple thing to determine and would
likely change from year to year. However, it should be ac-
knowledged that there may be instances in which carriers do
seek out longer hauls to make additional distance-related pro-
fits. If this tendency is pronounced, these profits would tend
to be bid away. This may provide a partial explanation for why
rates sometimes tend to taper with distance (i.e., the rate-
distance gradient tends to decrease with distance).

Multiple Committed Carriers

To this point much of the discussion has focused on the
effect of distance on committed carrier full-empty decision
making. In this section, differences are identified between
carriers which affect the net revenues and, hence, full-empty
decision making. These are incorporated into a model to ex-
plain the probability that committed carriers will move toward
their destination, j, full. In equation (1) the revenue and
cost elements have been specified that determine the net reve-
nue of an i-to-j movement. Each righthand side element will be
discussed in turn with regard to how and if it would be ex-
pected to differ across carriers.

The distribution of rates offered for i-to-j movements,

RATEij, is assumed to be identical across carriers. For the
empirical portion of this bulletin, this assumption is realis-
tic because the large majority of the movements examined are
regulated.6 This assumption is not crucial to the model, how-
ever. It could be relaxed at the cost of some additional com-

plication, but with no substantive changes in the model's im-

The full-empty running or distance-related cost differen-

tial, (FRCij ERCij) is also assumed to be invariant, or es-
sentially so, across carriers. As was previously argued, there
is little or no additional distance-related time associated
with operating full rather than empty. Therefore, additional
fuel and maintenance costs form the major share of this cost
component. As all of the carriers in the empirical portion of
this study operated similar equipment (full-size tractor-trail-
ers), differences in added (not total) fuel expenditures from
operating full or empty are likely to be negligible. Again,
this assumption makes no substantive changes in the model's

Differences among carriers are expected with respect to
direct search costs (DSC), pickup and unloading costs (PLU),
and the opportunity cost associated with a delayed arrival time

at j, (RATEjk) (DPRij). DSC and PLU involve time which may
differ from carrier to carrier according to the level of famil-
iarity with shippers in the area and loading/unloading facili-
ties and procedures. Carriers less familiar with an area may
pay premiums for brokerage services, motels, meals, loading
services and gate fees. Moreover, driver wage rates may differ
across carriers. If the carrier does not have an ICC authority
(i.e., a permit to haul regulated commodities interstate), a
premium or rent must be paid to a holder of an authority (a
trip or permanent lease). Finally, (RATEjk) (DPRij) would
differ across carriers if they anticipate acquiring different
outbound loads to different destinations (i.e., to different
k's) after j is reached. The time to search for and handle the
i-to-j load may differ across carriers which would affect DPR.


This section has two purposes. First, the type and col-
lection procedures for the data are briefly described. Second,
an overview of the trucking market serving Florida perishables
is presented. In some cases the data are employed to illus-
trate points which will relate to the development of the empir-
ical model in the following section.

Data Collection

Data were collected regarding the operations of truckers
hauling Florida-origin produce/ornamentals. This was done by
interviewing truckers as they left the state with such a load.
The types of information requested included:

1. carrier type, driver experience, and general
operating characteristics,
2. current outbound load type, acquisition method,
and trip characteristics,
3. load type, acquisition method, and trip charac-
teristics for the last inbound trip to Florida
(i.e., for the trip which brought them into the

The survey site was the outbound Florida Agricultural
Inspection Stations on U.S. 1-95 during the 1982-83 growing
season. Four survey periods were spread across the growing
season: November 11 and 12, January 26 and 27, March 28 and
29, and June 1 and 2. The rationale behind this spacing across
time was to gather observations during periods of growth (No-
vember), stability (March), and decline (January and June) in
shipment volumes (Figure 3). These seasonal swings in shipping
volumes are fairly regular, and shipment and freight rate lev-
els are known to be positively related (Beilock, Kohburger, and


This section has two purposes. First, the type and col-
lection procedures for the data are briefly described. Second,
an overview of the trucking market serving Florida perishables
is presented. In some cases the data are employed to illus-
trate points which will relate to the development of the empir-
ical model in the following section.

Data Collection

Data were collected regarding the operations of truckers
hauling Florida-origin produce/ornamentals. This was done by
interviewing truckers as they left the state with such a load.
The types of information requested included:

1. carrier type, driver experience, and general
operating characteristics,
2. current outbound load type, acquisition method,
and trip characteristics,
3. load type, acquisition method, and trip charac-
teristics for the last inbound trip to Florida
(i.e., for the trip which brought them into the

The survey site was the outbound Florida Agricultural
Inspection Stations on U.S. 1-95 during the 1982-83 growing
season. Four survey periods were spread across the growing
season: November 11 and 12, January 26 and 27, March 28 and
29, and June 1 and 2. The rationale behind this spacing across
time was to gather observations during periods of growth (No-
vember), stability (March), and decline (January and June) in
shipment volumes (Figure 3). These seasonal swings in shipping
volumes are fairly regular, and shipment and freight rate lev-
els are known to be positively related (Beilock, Kohburger, and


1200 -

900 -

- 800 -

700 -

600 -

.0 500 -
, i





Sept. Oct. Nov. Dec. Jan. Feb. March April May June July Aug.

Figure 3: Monthly Florida fresh fruit and vegetable shipments, Average 1976/77 to 1979/80.

During each survey period, the drivers of all trucks car-
rying exempt produce (essentially all produce except bananas)
or ornamentals and passing through the station between 6:00
P.M. and midnight (the high flow hours) were interviewed. In
all, 623 interviews were completed--112, 98, 198, and 215 in
November, January, March, and June, respectively. Only five
carriers refused to cooperate.

Truck Market Overview

Florida is second only to California as a producer and
exporter of perishables. Annually, between 150 and 200 thou-
sand truckload equivalents are shipped from Florida to the rest
of the nation and Canada. Approximately 90 percent of these
movements are by truck (U.S.D.A. and Strain and Beilock).
Virtually all of these trucks are full-sized tractor-trailers
with refrigeration units for temperature and humidity control.

The transportation system must be very flexible. This is
due to both the aforementioned seasonal swings and the con-
stantly changing mixture of products and origin and destination
points. Moreover, the system must be able to adapt to periodic
shocks or disruptions in the regular flow of freight due to
natural disasters such as freezes or other causes. Much of
this flexibility is realized through the participation of car-
riers which, much like migrant farm workers, serve those areas
then needing their services.

There are a large number of carriers involved in Florida
produce/ornamentals haulage. In 1980, the Florida Department
of Agriculture and Consumer Services (FDA&CS) recorded over
20,000 active carriers and over 200 transportation brokers.
The carriers may be divided into three groups or types: owner-
/operators, for-hire fleets, and private fleets (i.e., firms
hauling their own cargoes). Of the 623 truckers interviewed in
the 1982-83 season, 51 percent were owner/operators, 28 percent
were for-hire fleets, and 21 percent were private carriers.

The purchasers of the transportation services normally are
receivers (i.e., produce/ornamentals are usually shipped
F.O.B., origin). According to the FDA&CS, there are in excess
of 2,000 active receivers of Florida produce/ornamentals.
Approximately 60 percent of their loads are arranged through
brokers (Beilock and Fletcher).

Nonagricultural and processed agricultural products make
up the large majority of what is shipped into Florida.7 These
are subject to regulation. Therefore, for most hauls into the
state, carriers must hold an ICC authority (permit) to haul the
cargo, utilize the authority of another carrier under a leasing
agreement, own the cargo, haul it illegally, or enter the state
empty. In terms of total tonnage, significantly more comes
into the state than goes out (Kilmer, et.al.).8

A major assumption of the study was that on the inbound
journey to Florida, the carriers, for the most part, could be
considered to be committed. That is, the carriers specialize
in the outbound produce/ornamentals traffic, with inbound,
primarily nonagricultural commodities being of secondary impor-
tance. There are several reasons for asserting that this,
indeed, was the case. First over 90 percent of the respondents
had refrigerated trailers. These units are more expensive to
purchase than dry freight (i.e., unrefrigerated) trailers, are
heavier, and have reduced loading space due to thicker, insu-
lated walls and roofs.

Possession of a refrigerated unit, then, is a strong indi-
cation of specialization in the carriage of commodities requir-
ing refrigeration. That this commitment is primarily to ex-
empt refrigerated commodities (as opposed to regulated red
meats, bananas, and processed food products) is indicated by
the following:

1. Over two-thirds of the average respondents' time was
reported to be spent hauling exempt goods, and

2. Almost three-fourths of the respondents drove for
carriers not possessing any ICC authorities. This is
a remarkably low percentage considering that these
authorities cost only a few hundred dollars and that
the chances for rejection of an application for an
authority are remote.

Another piece of evidence in support of the contention
that the sample by and large was made up of carriers committed
to traveling to Florida relates to the manner in which the
sample was drawn. Drivers of vehicles exiting Florida with a
produce/ornamentals load were interviewed. Drivers of empty
vehicles were not interviewed. A carrier coming to Florida
with the certainty or strong expectation of acquiring a load
out of the state (i.e., a committed carrier) would be less
likely to leave the state empty than one whose primary load was
on the inbound leg of the journey (i.e., an uncommitted carri-
er). Therefore, the preponderance of the drivers of the empty
vehicles who were excluded from the sample were likely to have
been uncommitted carriers. Moreover, 35 percent of the drivers
interviewed had entered the state empty. By definition, these
must be committed carriers. Finally, 49 percent of the respon-
dents had their outbound loads arranged prior to entering Flo-
rida. This is a strong indication that the carrier was commit-
ted to the inbound trip.


The results of the conceptual model section suggests that
distance may be the only explanatory variable needed to explain
the proportion of committed carriers that will travel from i to
j with a load. This is true if truckers have homogeneous char-
acteristics (at least with regard to factors influencing their
costs) and that they operate in an environment of unchanging
market conditions. These assumptions do not adequately reflect
conditions in the Florida perishables trucking market. In this

section, variables are described which capture variations a-
cross carriers and time (Table 2).

Employing the previously hypothesized distance-rate and
distance-cost (FRC ERC) relationships as a justification,
distance (DIS) is employed to proxy expected net revenue E(P)
for hauls to Florida. It is hypothesized that E(P) is posi-
tively related to DIS because Florida is a net importer of
freight from all but four states east of the Mississippi River
(Kilmer, et al, p. 9). Therefore, the probability of a full
movement, P(Full), is expected to be positively related to
DIS. The additional explanatory variables to be described will
have the effect of shifting the E(P)-DIS curve upward or down-
ward, thereby influencing P(Full).

Market Conditions

The opportunity cost of delaying departure to Florida,
represented in equation (1) by (RATEjk)(DPR), may be proxied by
measures of overall market conditions. Capturing changes in
market conditions, however, is difficult without detailed data
on prices and their movements. In the present case, the south-
bound loads are for the most part regulated and somewhat resis-
tant to changes in rates over a few months. By contrast, the
northbound perishable loads picked up at the remote point (Flo-
rida) are unregulated. Rates for these commodities tend to ebb
and flow with the level of shipments (Beilock, et al, 1984),
and these changes affect the urgency of proceeding on from the
northern point (i) or, equivalently, the costs of delay,
(RATEjk) (DPR). The expectations about future rates are de-
rived from the seasonality of rates because of the high degree
of predictability in the seasonal nature of rates.

Produce shipments from Florida follow a fairly regular
seasonal pattern (Figure 3). Beginning in September, perisha-
bles shipments out of Florida slowly build to a peak around the
first of the year and then fall somewhat through February,

Table 2. The definition of variables included in Xn.

Variable Definition

Constant Unit vector

DIS One way mileage

DNOb DNO 1 if November, zero otherwise

DJAb DJA = 1 if January, zero otherwise

DMAb DMA = 1 if March, zero otherwise

PRIOR PRIOR 1 if load in Florida arranged
prior to arriving in the state, zero

FLE Ac FLE A 1 if fleet carrier with an ICC
authority, zero otherwise

OWN AC OWN A 1 if owner-operator with an ICC
authority, zero otherwise

OWN NAc OWN NA 1 if owner-operator without an
ICC authority, zero otherwise

PC Ac PC A = 1 if private carrier witn an ICC
authority, zero otherwise

PC NAc PC NA = 1 if private carrier without an
ICC authority, zero otherwise

FLH FLH = 1 if carrier based in Florida, zero

ORH ORH = 1 if carrier based in origin state
of southbound load, zero otherwise

aThe intecept corresponds to a fleet carrier whose home is not
located at either end of the southbound trip, who has no ICC
authority and who hauls during the month of June.
bDJU, the corresponding binary variable for June is the omitted
cFLE NA the corresponding binary variable for fleets without
an ICC authority is the omitted category.

stabilizing at a moderate level (around 700 million pounds per
month) from February through March or mid-April. In late March
or early April shipments begin an increase to around 1,200
million pounds per month by May and fall to almost zero by
August. The survey data were gathered in late November, mid-
January, late March, and early June. Urgency to proceed back
to Florida for a load is expected to be highest in late May and
early June when shipments are near their annual peak and begin-
ning to fall rapidly, and lowest in November and March when
shipments are either building or when an increase is expected.
To proxy the effects of these differing market conditions,
binary variables are specified for the first three survey
months (DNO, DJA, DMA) and are expected to have positive signs
(relative to June (DJU) which is represented in the intercept).

Carrier Characteristics

This section describes the variables used to capture the
heterogeneous characteristics among carriers that influence
costs. To capture differences in urgency among carriers to
depart for Florida a binary variable (PRIOR) is specified to
indicate if the load which was picked up in Florida had been
arranged prior to entering the state. The probability of being
full on the southbound movement is expected to be negatively
related to PRIOR. Searching for, picking up, and dropping off
an inbound load inevitably consumes time which a carrier having
a previous commitment may not be able to afford.

Several aspects of the carrier's status are of interest as
they influence DSC and PLU. First, if the carrier has an ICC
authority to carry regulated goods, then there is no need to
work through an intermediary by trip leasing or to embark on a
search for hard-to-find unregulated commodities. Carriers with
such an authority would be expected to have lower DSC. There-
fore, they would be more likely to secure a load to Florida
than those without ICC authorization.

Second, carriers may differ with respect to DSC depending
upon their status as an owner-operator (OWN), a private carrier
(PC), or a fleet carrier (FLE). Due to their small size,
owner-operators generally rely on brokers of regulated carriers
to which they may lease their services in order to secure loads
(Taff). Private carriers historically have been prohibited
from carrying regulated freight. While the regulatory picture
in this regard began to change with the Motor Carrier Act of
1980, many private carriers continue to limit operations to
their own goods and to for-hire exempt agricultural products.
For-hire fleets form thepbackbone of regulated freight truck-
ing. They are the most likely to have the broadly defined
authorities and the sales and dispatching services to facili-
tate securing and handling regulated freight (Wyckoff and Mais-
ter, pp.xxxii-xxxiv)

To capture the effects of differences in DSC and PLU be-
tween the regulatory status-carrier type groups, six binary
variables are specified:
FLEA, FLE_NA, OWNA, OWNNA, PC_A, and PC_NA .10 Fleet car-
riers without authority (FLENA) are the omitted category for
the purpose of estimation.
FLE A is expected to be positively related to full movement
and PC NA to be negatively related. The remaining groups are
expected to fall in between.

Finally, to proxy carrier familiarity with the markets, a
binary variable is specified to indicate if the carrier is
based in Florida (FLH) or in the origin state of the movement
into Florida (ORH). It is hypothesized that carriers based at
one end of the run are likely to have lower DSC to arrange a
load than those based in a third state, and hence more likley
to search for loads, ceteris paribus. With regards to the
southbound movement, the probability of acquiring a load would
be enhanced only if this increased market familiarity lowered
DSC proportionately for both the north and southbound runs or
disproportionately favored the southbound run. If, on the
other hand, the carrier's special knowledge increased for

northbound load acquisitions but not for southbound load acqui-
sitions, the probability of acquiring a southbound load would
be lower, ceteris paribus.


Of interest is estimating the probability of being full,11
P(FULL), or the full/empty odds ratio, P(FULL)/P(EMPTY). Fol-
lowing Maddala (pp. 22-25), we assume that P(FULL)/P(EMPTY) can
be represented by the following equation:

(7) P(FULL)/P(EMPTY)i = Xio + Ui.

Xi is the row vector of exogenous variables (Table 2); 8 is a
column vector of coefficients, and U. is the error term.

Unfortunately, P(FULL)/P(EMPTY) is not observable. It is
replaced by the binary variable FULLi equal to one if the car-
rier travelled to Florida with a load, and zero if empty.
Assuming that Ui is represented by a logistic cumulative dis-
tribution,12 the likelihood function is:

n I (1 Full ) exp(XiB) (Fulli)
(8) L = n ( ) Fu ( +exp
,=1 i+exp(X.) 1+exp(X 8)

where L represents the likelihood function, n is the number of
observations, and exp is an operator raising e to the(X i) pow-

The Newton-Raphson nonlinear optimization method is used
to maximize the log of equation (8) in order to obtain maximum
likelihood estimates for the coefficients (SAS, pp. 181-202).
Initial parameter estimates are arbitrarily set to zero. Con-
vergence to an optimum is assumed when the difference in -2 log

likelihood between two successive interations is less than
.025. The estimates are consistent and asymptotically effi-


The log likelihood ratio for the estimated equation is
significant at the .01 level (Table 3). The percentage of
carriers correctly categorized by the model is 85.7. Moreover,
the signs and magnitudes of the estimated parameters are gener-
13 14
ally in accordance with expectations.314

As hypothesized, DIS is positively associated with full
movements into Florida (the estimated parameter associated with
DIS equals .0038, Table 3). This suggests that a RATE /8DIS is
greater than a (FRC ERC)/9DIS. Given that Florida is a net
importer of freight from almost every state east of the Missis-
sippi River (Kilmer, et al, p. 9), it is not surprising that
inbound rates would tend to be above the differential between
full and empty running costs.

June was the omitted category for the monthly binary vari-
ables. The positive signs of the estimated parameters for the
remaining months indicate that carriers are more likely to
acquire southbound loads into the state than they were in June.
This is consistent with the expectation that in June there
would be the most urgency to return to Florida due to
collapsing shipment volumes and rates on loads from the state

(i.e. (RATEjk)(DPR) would be negative and large). The relative
magnitudes of the estimated parameters are also in accordance
with expectations, with the estimated parameters associated
with March and November being larger than that for January.15

The impact of market conditions on full/empty movement
decisions is depicted in Figure 4. March is at the beginning
of the rapid rise in produce shipments and rates which occur at
the end of the Florida crop season (Figure 3). At that time,
Delays in getting to Florida are likely to result in the carri-

Table 3. The results of a logit analysis of full/empty
southward movements of motor carriers of Florida
produce and ornamentals, 1982-1983.

Asymptotic Change in
Variable Coefficient standard errors probability

Intercept -2.9482***b (.5652)

DIS .0038*** (.0004) .0006008

DNO .9633*** (.3614) .1514

DJA .6521* (.3466) .1025

DMA 1.1355*** (.2855) .1785

PRIOR -.8311*** (.2373) -.1306

FLE A 1.4813*** (.5098) .2328

OWN A .2501 (.5667) .03932

OWN NA -.4518** (.3709) -.0710

PCA -.8025 (.8350) -.1261

PCNA -.8365* (.4300) -.1315

FLH -.4034 (.2896) -.06340

ORH -.4316 (.2262) -.06785

Log likelihood ratio chi square
(12 degrees of freedom) 230.73***c

Percent of carriers correctly
categorized by the model 85.7

Number of observations 574d

***Significant at the .01 level.
**Significant at the .05 level.
*Significant at the .10 level.
aCalculated at the mean levels of the independent variables.
See footnote 11.
bSignificance levels are based on the Wald Statistic.
cThe null hypothesis assumes a model with a constant only.
dForty-nine observations deleted due to incomplete information.

...... March



/ /

.* / /

* /

*. /


- I I I I

0 200 400

600 700 800 1000 1200 1600 1800


Figure 4: Probability of full movement by month and







0.3 -


er finding more favorable conditions than if the inbound trip
had been speedier. The reverse is true for June, which comes
just after the season-end shipping peak. Delays in this month
result in the carrier finding fewer loads, generally at lower
rates, than if he had arrived in Florida sooner. On average,
the probability of observing an average carrier with a south-
bound load is greater than .5 percent for inbound hauls of 800
miles (Figure 4). In June, however, the minimum linehaul dis-
tance to observe at least half of the carriers with southbound
loads is 980 miles. These results may be due in part to season-
al variations in southbound shipments. The authors are not
aware of such variations. Moreover, if they exist, they are
likely to be slight compared to the variations in produce ship-

Prior arrangement of an outbound load from Florida sharply
reduces the probability of hauling a load into the state. The
estimated parameter associated with PRIOR (-.8311) is highly
significant. At mean levels for the independent variables, a
carrier which has a prearranged load in Florida has a .13 lower
probability than other carriers to enter Florida with a full
load, ceteris paribus.

The estimated parameters associated with the regulatory
status carrier type groupings indicate the following:

1. fleet carriers with authority have the highest
probability of securing a southbound load, while
private carriers without authority have the lowest
probability, ceteris paribus (Figure 5), and
2. possession of an authority appears to be of sig-
nificant value in attaining a southbound load for
fleet carriers, but not for owner-operators and
private carriers (Table 3).

These results must be interpreted with care because at least
two effects are operating at once: differences in organization
between carrier types and authority ownership or nonownership.
Two additional possible effects which add further complication
are organizational differences between firms within a carrier

1.0 -


/ /
0.8 /

S 0.7 .

o /
." /
.- 0.6 /
- .
LL /
0 0.5 -
>. / /
I- /
- / /
S 0.-4 /
O /
0 /
- /
0.3 /

0.2 /
/ ........ Fleet with authority
0.1 Average
-- Private carrier
without authority
0.0 II I I

0 200 400 600 800 1000 1200 1400 1600 1800


Figure 5: Probability of full movement by carrier type,
ICC authority.



type with and without an authority, and qualitative differences
in authorities between carrier types.

Fleet carriers concentrate on for-hire carriage and cus-
tomarily employ sales staffs to generate business in both di-
rections for its vehicles. Private carriers, on the other
hand, primarily haul their own goods. Theic fleets are gener-
ally ancillary to their sales and production departments, with
much less emphasis placed on securing for-hire carriage. Own-
er-operators fall somewhere between the two other types. They
concentrate on for-hire carriage, but lack the organization
necessary to maintain close contacts with a wide range of po-
tential shippers. It is consistent with the above that fleet
carriers possessing the added benefit of an authority were the
most likely to secure southbound loads, while private carriers
without authorities were the least, ceteris paribus.

It is also consistent that only fleet carriers appear to
derive significant benefits (vis a vis southbound loads) from
holding authorities. Their aforementioned organizational
strengths make it easier for them to make and maintain the
contacts with shippers for which (the carriers') authorities
allow them to haul. Moreover these same organizational
strengths give them advantages when seeking authorities. Own-
er-operators and private carriers are likely to seek authori-
ties limited to a few commodities, shippers, or areas. Fleet
carriers, on the other hand, are more likely to seek out and
secure authorities allowing for very broad latitudes with re-
spect to permissible areas, shippers, and commodities.

It cannot be concluded absolutely, however, that authori-
ties in themselves significantly improve the probabilities for
southbound loads of fleet carriers. It seems reasonable that
fleet carriers with authorities may be more aggressive in seek-
ing loads than those without authorities. That is, possession
of an authority may be a proxy for the aggressiveness of fleet
carrier managements, rather than a pure indicator of the load
acquisition value of an authority. Almost surely, the results
represent some mixture of the two effects. However it should

be stressed that even if a very large share of the observed
effect is due to organizational differences, it still can be
said that aggressive firms believe that authorities are of

The estimated parameters associated with the final two
variables, FLH and ORH, were insignificantly different from
zero at the .10 level. This suggests that being based at one
endpoint of the inbound movement does not impact materially on


In this bulletin the determinants of full/empty movements
for motor carriers have been examined. It was hypothesized
that these decisions would depend upon the differential between
the rates received and costs associated with the load, the
urgency of proceeding to the destination (the opportunity cost
of delay), carrier regulatory status, firm characteristics, and
carrier familiarity with the market. An empirical model was
developed and estimated employing data on movements into Flori-
da by carriers which haul produce out of Florida. As the de-
pendent variable (moving full or empty into Florida) was quali-
tative, logit analysis was employed. The equation is highly
significant and correctly categorizes 86 percent of the carri-
ers sampled.

Several important conclusions can be drawn from the study.
First, the results suggest that rates rise more quickly with
distance than does the increment in costs from running full
rather than empty. This result is reasonable in a region such
as Florida, for which more freight is incoming than outgoing.
This suggests a direction for future research. If there are
differences between net importing and net exporting regions (as
is hypothesized), then the rate-distance gradients, both in-
bound and outbound, should be examined. It may be that along
routes, such as the southbound route examined in this study,

rates taper (increase at a decreasing pace) with greater dis-
tance as a larger and larger proportion of carriers committed
to travel along the route seek loads. If true, this would
provide an additional rationale to those offered by Bressler
and King (pp. 108-16) for explaining the phenomenon of tapering

Another conclusion of the study is that expected varia-
tions in rates at remote points have an impact upon carrier
full/empty movement decisions. In particular, expected rises
(Falls) In rates at remote points lower (increase) the oppor-
tunity cost of the delays inherent in seeking and handling a
load to that point. This, in turn, has a significant impact
upon carriers' full/empty movement decisions, suggesting a very
high level of carrier sophistication in the search process. In
future research it would be of interest to test for degrees of
risk averseness as this would have implications regarding the
value of stable, possibly regulated, rates.

The final major implication of the study is that ownership
of an ICC authority appears to be an important determinant of
southbound full/empty movements for fleet carriers. This
strongly suggests that, despite the reforms of the Motor Carri-
er Act of 1980, the interstate regulatory structure continues
to contribute to unnecessary empty movements. The magnitude of
the added costs to the agricultural transportation sector of
these regulation-forced empty movements can only be speculated
at, but that they exist is almost a certainty.

Strictly speaking, all conclusions can be applied only to
the Florida routes studied. However, the hypotheses were gen-
erated from a non-Florida specific conceptual model. Thus,
these results are hypothesized to be valid for all net import-
ing regions. This hypothesis needs further testing before the
conclusions can be generalized with complete confidence.


1. This is the transportation analog to peak load pricing.
The heavier direction of traffic must bear the capacity
costs for the transportation services. A full explanation
of this is presented in Robbins and Reed. It should,
perhaps, be noted that this is a long run solution. In
the short run a carrier may travel empty in one direction
without being able to recoup the full costs of the round
trip on the return leg of the journey.

2. A similar situation exists for a profit maximizing farmer
who will harvest his crop only if the revenue exceeds the
variable costs of harvesting. The return does not need to
be sufficient to cover the total cost of already completed
task, such as planting, weeding, pest control, and irriga-
tion, for these are sunk or committed costs.

3. With risk aversion there would be some discounting of the
value of the expected, but uncertain return (Pij) and the
reverse for risk lovers. This could easily be incorporat-
ed into the model by inflating or deflating (Pij) by a
factor to account for deviations from risk neutrality.
If, as is generally the case, risk aversions is assumed to
exist and to be reasonably consistent across carriers,
then the analysis would not be materially altered.

4. The expected rate is that which the carrier anticipates to
be in force at a given time and place. A model is not
developed to explain how these expectations are arrived
at. This is not necessary, however, as it is the expected
movement or change in the rates that is of importance. We
argue that the regularity of the seasonal pattern for
produce/ornamentals shipments and rates allows carriers to
base their expectations of rate movements on the time of

5. For committed carriers hauling into net exporting regions,
net revenue would not be expected to change with
distance (i.e.,(P)/3(DIS) = 0).

6. Many carriers do have exceptions or carrier-specific rates
filed with the Interstate Commerce Commission, and the
Motor Carrier Act of 1980 does allow for some flexibility
in rates. However, it appears, from informal conversa-
tions with carriers, that the majority of regulated
freight is still charged in accordance with rates deter-
mined collectively by rate bureaus.

7. Two hundred and ninety-nine of the reported four hundred
and twenty-four inbound loads were subject to ICC regula-

8. Kilmer, et. al. found that Florida is a net importer of
freight from almost every state. This has been corrobo-
rated in informal conversations with representatives of
the Florida Trucking Association and the Seaboard System
Railroad. They estimated between a 6-to-l and 9-to-1
freight imbalance inbound for the state.

9. Since the late 1970's, the ICC has approved almost all
applications for authority (Harper, 1970).

10. An alternative specification which would yield similar
results follows:

P(FULL)/P(EMPTY) = B0 + B1 OWN + B2PC + B A + B OWN_A

+ B5 PC_A + (other variables as shown
in Table 2).

11. Partial loads are not very common. Moreover, they may be
treated as another type of load, just as a load of steel
is different from a load of pillows. Whether you identify
it as a partial or a full load, the vehicle is being hired
to perform a given movement. Vehicles have two basic
limitations space and weight. A full load of steel
might take up very little space, while a full load of
pillows might weigh very little. Differentiation is not
really crucial if partial loads are viewed as another type
of load.

12. An alternative to logit analysis is probit. Formally, the
choice between logit and probit analyses depends upon the
assumptions made regarding the distribution of the depen-
dent variable. With logit analysis a logistic distribu-
tion is assumed, while a normal distribution is assumed
for probit analysis. Because these distributions are
similar (a logit distribution has somewhat fatter tails
than a normal distribution), as a practical matter the
selection of one or the other generally has little effect
on the results (Capps and Kramer). Given this, and the
ready availability of software to perform logit analysis,
it was chosen.

13. A heteroskedasticity test was performed because of the
cross-sectional nature of the data. The heteroskedastic
nature of the error term was assumed to

be E(u2) = (2)(DIS)-6. A Lagrangian multiplier test (LM2)
was used (Davidson and MacKinnon). The null hypothesis of
a homoskedastic error term could not be rejected.

14. With logit analysis the estimated parameters do not di-
rectly show the change in probability (in this case, the
probability of having a load) associated with a one unit
change in the independent variable. This is conditional

on the rest of the model's parameters and variable values.
The formula to calculate this follows:

SPRB (( EXP(XB))/(1 + EXP(XB))) B .

For mean values of the independent variables, these values
are presented in the rightmost column of Table 3.

15. Aberrations in the movement of goods into Florida may
contribute to this pattern. However, no such fluctuations
are known. Moreover, any fluctuations in southbound ship-
ments of largely manufactured dry goods would be mild
compared to those for Florida produce. Therefore, the
explanation given in the text for the seasonal pattern of
observed full/empty movements seems the most reasonable.


Baesmann, R. and A. Daughety "Complementarity in Production
and Empty Backhaul in Transportation." Presented at
Western Economics Association, 1977.

Beilock, Richard and G. Fletcher "Exempt Agricultural
Commodity Haulers in Florida." Proceedings of the
Transportation Research Forum. 24,1(1983):444-50.

Beilock, R., J. Kohburger, and J. Morgan "Explaining Truck
Rate Variations for Produce Shipped to the Northeast."
JNAEC 13,4 (1984) 25-32.

Bressler, R. and R. King. Markets, Prices and Interregional
Trade. New York, John Wiley & Sons, Inc., 1978.

Boles, P. Owner-Operator Costs of Hauling Fresh Fruits and
Vegetables in Refrigerated Trucks. USDA, ESCS report No.
82, 1980.

Capps, 0. and R. Kramer "Discriminating Among Alternative
Qualitative Choice Models." Unpublished manuscript,
Department of Agricultural Economics, Virginia Tech, 1983.

Davidson, Russell and James G. MacKinnon. "Convenient
Specification Tests for Logit and Probit Models." J.
Econometrics 25(1984):241-62.

Harper, D. "The Federal Motor Carrier Act of 1980: Review and
Analysis." Transportation Journal. 20,2(1980) :5-33.

Jara-Diaz, S. "Transportation Production and Cost
Functions." Transportation Science 16,4(1982):522-539.

Johnson, M. "Current and Developing Issues in Interregional
Competition and Agricultural Transportation." SJAE

Kilmer, R., H. Ramirez and F. Stegelin Economic Impact of
Empty Backhauls in Florida Fresh Fruit and Vegetable Truck
Transportation. Florida Experiment Station Technical
Bulletin No. 833, 1983.

Klindworth, K. and E. Brooks. Shipping Alternatives for
Moving Florida Fruits and Vegetables to Eastern and
Midwestern Markets. U.S.D.A., Office of Transportation,

Maddala, G.S. Limited-Dependent and Qualitative Variables Ln
Econometrics. Cambridge: Cambridge University Press, 1983.

Paxson, D. "Motor Carrier Deregulation and the Opportunities
for Reducing Empty Truck Mileage." Association of American
Railroads Staff Paper 79-2, 1979.

Robbins, L. and M. Reed. "Rail Deregulation, Efficiency, and
Peak-Load Pricing: A Theoretical Analysis." University of
Kentucky Department of Agricultural Economics Staff Paper
#133, 1982.

SAS Institute Inc. SUGI Supplemental Library User's Guide.
1983 Edition. Cary, N.C.: SAS Institute Inc., 1983.

Stigler, G. "The Economics of Information." JPE

Strain, R. and R. Beilock. "Truck Shipments of Florida
Ornamentals to the U.S. and Canada." Proceedings of the
Florida State Horticultural Society 95(1982): 181-184.

Taff, C. "A Study of Truck Brokers of Agricultural Commodities
Exempt from Economic Regulation." Transportation Journal
19,3(1979): 5-15.

U.S. Department of Agriculture. "Fruit and Vegetable Rate
Report." A.M.S., 1979-1981 (weekly).

Wilson, G. Economic Analysis of Intercity Freight
Transportation. Bloomington: Indiana University Press,

Wyckoff, D. and D. Maister The Motor Carrier Industry.
Lexington: Lexington Books, 1977.


All programs and related activities sponsored or assisted by the Florida
Agricultural Experiment Station are open to all persons regardless of race,
color, national origin, age, sex, or handicap.

This publication was produced at a cost of $1,186.00 or $1.186 per copy
to report technical information to the professional and scientific com-
munity concerning full-empty truck movements into Florida.

ISSN 0096-607X

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs