Printer Friendly

Private fleet use: a transaction cost model.

Private trucking fleets are an important part of the United States transportation system. Although private fleet use was expected to decrease with deregulation, recent surveys indicate that hundreds of thousands of private fleet units remain in service, and the number may even be increasing.(1) Experts estimate that up to 50 percent of all domestic trucking is currently performed by private fleets.(2) Private fleets vary in size from one or two power units to 500 or more, and are found in many different industries.(3) Shippers and receivers continue to view private trucking as an important alternative in the transportation choice process.

At the same time, the decision processes surrounding private fleet use have not been studied in any depth. The logistics literature contains extensive research concerning the choice of transportation mode and carrier; the studies in such areas have been conducted over the past twenty years, encompassing issues of cost, customer service, and selection procedures in many settings. However, this research has not recognized the private fleet as a separate alternative, although it seems likely that the choice to own and operate a private fleet is based on very different concerns than typical mode or carrier selection decisions.

The current study conceptualized private fleet use in terms of the outsourcing or "make-or-buy" decision. The make-or-buy issue has been discussed for many years in the purchasing and operations management literature, and many factors appear to influence this process.(4) During the past fifteen years economists have suggested that transaction costs are a major determinant as to why and how outsourcing occurs. This theory, known as Transaction Cost Analysis (TCA), posits that a company buys goods and services on the open market unless certain specific conditions characterize the buyer-seller relationship. Where such conditions exist, the firm tends to "make" rather than buy, the entities such as private fleets evolve. TCA has been tested empirically in studying distribution channel choices and industrial product make-or-buy decisions.(5)

This article reports on a test of the influence of TCA in the context of private fleet use. The article is divided into five sections: The first outlines the background of the research problem including the importance of the private fleet decision and the role of the private fleet in the firm's transportation function. The second section reviews the economics and logistics literature relevant to the choice of a private fleet alternative. Purchasing, TCA, and previous carrier selection literatures are reviewed in this section. The third section details the design and execution of the research study, including questionnaire design, pretesting, and sample selection. The fourth section discusses the results of the survey and the test of the TCA based model. The fifth section considers the implications and limitations of the survey results and suggests areas for further research.


Logistics researchers have recognized the large size of the private fleet sector. In 1990, Anderson and Gillies(6) stated that "private carriage still accounts for almost half of intercity trucking." La Londe et al.(7) found that approximately 15 percent of all raw material and finished goods tonnage are moved by private motor carrier. Delaney(8) has estimated that the cost of private and proprietary trucking was $84 billion in 1989 versus an estimated $72 billion expenditure for public/for-hire trucking.

In spite of the economic significance of the private fleet alternative, prior studies have centered around 1) the choice between rail and truck, and 2) the selection of specific carriers within mode. McGinnis(9) summarizes many of the studies performed to determine the factors involved in choice of transportation mode and carrier. These studies have consistently found that both service and cost are important selection factors, with service generally rated somewhat more important than cost. Only the study done by Jones(10) addressed the private carriage choice explicitly. In his 1975 study, Jones found that service was the primary motivator for use of a private fleet.

Outsourcing of logistics in general has also drawn increased interest in the past few years. Cavinato(11) suggested that firms may be reducing private fleets and other logistics assets primarily for reasons of cost and service. Fernie(12) found that retailers still doing their own distribution believed that they provided better service at a lower price than outside contractors. Fernie also found that organizational history and inertia played a large part in the continued operation of private fleets.

The economics literature suggests another approach to the issue of private fleet use. Building on the work of Coase,(13) Oliver Williamson has expounded the theory of "efficient boundaries."(14) Williamson theorized that companies internalize business functions to minimize the transaction costs incurred in such functions. Such transaction costs arise from three main sources: Transaction-specific assets develop when either the shipper or the carrier makes an investment which is worthless if applied in any other context. Beier's(15) driver-based learning is one example of such assets. A driver who picks up or delivers from a customer daily will be familiar with the customer's procedures and shipping or receiving area. This knowledge increases the driver's efficiency for this specific customer but is not transferrable to other customers. The risk of losing such knowledge represents an opportunity cost, and the shipper may seek to minimize this risk/cost by vertical integration--using the private fleet. Other types of specific assets include equipment customized for or dedicated to single customers and facilities built to serve a particular customer location. Uncertainty may encourage renegotiation of shipper/carrier arrangements with a likely increase in costs due to carrier knowledge of the shipping situation. The shipper can eliminate such opportunism by substituting private fleet service. Williamson differentiates between "external uncertainty," which is generated by volatility in the environment, and "internal uncertainty," which results when the firm cannot easily assess employee and/or vendor performance. Frequency of shipment can lead to potential economies of scale. The shipper can enjoy such economies best by serving the customer with the private fleet. TCA predicts the use of private carriage where there are high levels of transaction-specific assets, uncertainty, and frequency.

Students of purchasing and industrial economics have suggested that organizational and situational factors are also important in the make-or-buy decision. Williamson proposed that larger companies are more likely to internalize functions, irrespective of other conditions.(16) Scherer et al. summarized previous theory and claimed that the structure of the downstream market is important to the integration decision.(17) Lee and Dobler(18) and Zenz(19) list numerous factors which should be considered in the make-or-buy decision. Some of these concerns, such as specialized supplier know-how and supplier reliability, fit within the TCA framework. Other factors such as plant utilization and human resource issues are not within the purview of TCA. Finally, Speh and Blomquist found that specific "trigger events" may result in a make-or-buy analysis.(20) Although Speh and Blomquist dealt with warehousing, it seems likely that particular events may also affect the choice to use private fleets.

The modeling technique and data gathered in this study allowed the simultaneous testing of the importance of TCA factors, company size, and market structure in the private fleet decision. However, the present work does not address the influence of situational factors or specific industry effects.(21) Inclusion of situational specifics would seem to require detailed case studies, and limited sample size precluded industry-specific analysis of fleet usage.


A survey was designed to test the applicability of TCA to the use of private carriage. The survey was pretested on a limited number of shippers and carriers. Questions were formulated to measure shipper perceptions of specific assets, uncertainty, and frequency associated with shipping to each firm's largest customer. These questions were adapted from prior TCA research in the areas of sales management and distribution channels.(22) The specific items pertaining to asset specificity, uncertainty, and frequency can be found in Table 1.

Surveys were sent to logistics professionals representing 488 organizations in the consumer goods, chemical, pharmaceutical, and automotive industries. One follow-up letter was sent four weeks after the initial mailing. One hundred forty-seven responses were received for an overall response rate of 30.1 percent. After eliminating surveys with missing data, 138 firms formed the data base for this study. Non-response bias was checked by comparing early and late responses, as outlined in Armstrong and Overton.(23) The results from the two groups were not statistically different, indicating that non-response bias is probably not a problem.


Since the study was concerned with the applicability of TCA, respondents were asked about the use of the private fleet to service a specific shipping transaction. The transaction chosen was the shipment of finished goods to the firm's largest customer. The percentage of total sales accounted for by this largest customer varied from 1 percent to 100 percent, with a median percentage of 9 percent.

Two dependent variables were considered. The first dependent variable was an indicator variable for any use of the private fleet. This variable was coded as 1 if the private fleet hauled any shipments to the primary customer and 0 if there was no private fleet use. Of the 138 study firms, 28 (20 percent) indicated some use of the private fleet to service their major customer.

The second dependent variable was the percent of this customer's shipments actually handled by the private fleet. The percentage varied from 4 percent to 100 percent. Table 2 indicates the distribution of private fleet use by industry and in total. Firms with private fleets seem to use both private and for-hire carriers to service their largest customer. Note that this variable is defined in terms of shipments, as advocated by Beier.(24) Defining this variable in terms of weight resulted in a similar distribution of private fleet usage. Table 2 indicates significant differences in private fleet use by industry. However, the total sample was utilized for testing and model building because of the relatively small sample size.



Principal components factor analysis was used to combine the multiple questions concerning specific assets and uncertainty. Although the number of questions was small, there were two reasons to combine individual items. First, an examination of the correlation matrix among the items suggested that multiple items might represent the same construct. Second, TCA suggests that both asset specificity and uncertainty have definite subcategories, as outlined in the literature review above. Principal components analysis was used to determine which categories might be represented in the current study. Two forms of specific assets were indicated. The first factor is dominated by questions concerning the cost and transition time of any change in carriers. This factor seems to be "human asset specificity," or "learning by doing," as suggested by Williamson.(25) The second factor represents the specific assets involved in customizing service for this customer. This construct may correspond to Williamson's "dedicated assets."

Uncertainty also appeared to have two facets. The first factor corresponded to external uncertainty(26) and was associated with variation in the size and timing of customer orders, as well as the difficulty in forecasting customer orders. The second factor reflected internal uncertainty(27) and was associated with shippers' concern as to their own ability to judge the current carrier's performance. Table 3 shows the factor loadings associated with the asset specificity and uncertainty variables. It should be noted that the factor structure is not pronounced in the analysis of the uncertainty items as in asset specificity. Since frequency was represented by a single item, principal components analysis was not required for this variable.
Table 3. Factor Loadings Associated with Specific Assets and
a. Specific Assets
 Human Dedicated
Question Specificity Assets
Service customization .115 .983
Cost of carrier change .898 -.015
Transition time of change .824 .269
b. Uncertainty
 External Internal
Question Uncertainty Uncertainty
Variation in order size .769 -.074
Variation in time between .497 .565
Forecasting difficulty .767 .104
Evaluation of carrier -.110 .901

One way to test the importance of TCA variables is to compare fleet users and non-users. If TCA is correct, the two groups should differ on the dimensions of asset specificity, uncertainty, and frequency. Table 4 shows the mean values of asset specificity, uncertainty, and frequency for users and non-users, as well as the results of t-tests to determine if the two groups are statistically different. The values and test results of size and market structure variables are also listed.

The importance of specific assets is strongly supported by the survey results. If a company perceives that changing carriers will be an expensive or lengthy process the company will tend to use the private fleet to service the customer. On the other hand, the need to customize service does not seem to drive the use of private carriage. Apparently logistics managers see private fleets as a way to reduce transaction cost and risk rather than as a means of providing specialized service. These results are supportive of one of the key tenets of TCA.

Findings concerning the role of uncertainty in fleet use are somewhat consistent with TCA. Companies currently utilizing private fleets are more confident of their ability to evaluate the current carrier's performance than companies using external suppliers. TCA would suggest that companies adopted private fleets partly to make the evaluation task easier. On the other hand, there is no significant difference in external uncertainty between users and nonusers. Relative frequency of transaction also does not differ statistically between the two groups, counter to the predictions of TCA.

Finally, both size and market structure appear to influence private fleet use. Private fleet users are statistically larger in terms of TABULAR DATA OMITTED shipment counts, although the difference is significant only at the .09 level. Perhaps only larger companies have the resources and internal demand to operate cost-effective private fleets. Interestingly, companies whose primary markets are manufacturers and wholesalers are less likely to use private fleets to service their largest customer, while companies serving small retailers are more likely to be private fleet users.


Respondents also provided information on the percentage of the customer's shipments delivered by the private fleet. TCA considerations suggest that heavier fleet users might have shipper-customer relationships characterized by higher levels of specificity, uncertainty, and frequency. Therefore, a multiple regression model was constructed based on data from the private fleet users. The dependent variable was the percentage of total shipments carried by the private fleet. The independent variables were the two variables representing asset specificity, the two variables representing uncertainty, and variables relating to frequency, firm size, and type of market served. The initial regression resulted in only two variables with coefficients different than 0, based on individual t-tests. After eliminating statistically nonsignificant variables, a fairly parsimonious model of private fleet use emerged. Indeed, the model below was superior to the original seven variable model, based on the F-statistic and adjusted |R.sup.2~.

% Shpts on Fleet = .545 (Human Spect) + .384 (%Small Retail) + 43.5

Based on the standard F test, the overall model is highly significant; the probability of all coefficients being 0 is less than 1 in 1,000 (p |is less than~ .0002). Also, the coefficients of both asset specificity (Human Spec) and %Small Retail are highly significant. For specificity, the probability of the coefficient being 0 is .016. For % small retailers, the probability of the coefficient being 0 is .001. The adjusted |R.sup.2~ of the model is .475, indicating that the variables in the model account for nearly half the variation in the data. The statistics for model evaluation are summarized in Table 5.
Table 5. Summary of Model Statistics
Quantity Value T-value (p-value)
Overall model 12.33 Not applicable .0002
significance --
F statistic
Model fit -- .475 Not applicable Not applicable
adjusted R2
Coefficient of .384 2.61 .0157
asset specificity
Coefficient of .545 3.71 .0012
% small retail

The model indicates that human or learning curve assets are an important determinant in how much the private fleet is utilized. If a shipper perceives that carrier changes would be costly or require significant carrier retraining, the shipper will increase the use of private fleet. Similarly, as the percent of small retailer weight in the shipper's customer base increases, the percent of private fleet use increases for major customers. The latter result may point to a two-tier service system, where the private fleet handles large customers and small customers are serviced by third-party transportation. The large size of the constant in the above model is also interesting. The large constant implies that shippers will frequently choose to handle a sizable portion of the customer's needs on the private fleet, rather than a small amount. This finding is consistent with Table 2, where the private fleet handled 30 percent or more of the customer's requirements in 70 percent of the cases.


The results reported above indicate that private fleet use is influenced by both transaction cost and market considerations. If a shipper perceives that carrier changes will result in increased cost and decreased service during the transition, the shipper will use the private fleet to minimize the chance of such changes. A major reason for the existence of private fleets may be to reduce risks associated with dependence on third parties. The empirical data support this risk reduction rationale. Shippers believe that carriers accumulate transaction-specific assets through mechanisms such as the learning curve. Such assets may give the carrier the power to "hold up" the shipper in subsequent negotiations. To forestall such opportunism, shippers will substitute private fleet service.

The market served by the shipper also appears to affect the propensity to use private fleet. In particular, shippers selling to small retailers are more likely to use private fleets for their largest customers. On the other hand, companies serving manufacturers and wholesalers are less likely to use private fleets for their large customers. Possibly manufacturers and wholesalers are large enough that they have their own fleets or can command large discounts from common and contract carriers. Alternatively, shippers may give large customers special attention when most of their market is smaller retail establishments.

The primary importance of this study to the logistics community is in the emerging area of third-party logistics. Logistics companies hoping to replace corporate fleets should realize that such fleets are often used to reduce risk. Third-party carriers have to demonstrate that they will not take unfair advantage of their knowledge of shipper needs and operations. One possibility is the upfront use of long-term contracts which will protect shippers from arbitrary increases. Some of the growth in contract carriage may result from shippers' willingness to substitute contract trucking for private fleet service.

This study also provides some explanation for the slow growth of third-party logistics in the United States. Third parties often point out that they can provide services more cheaply, save shippers capital, increase return on assets, etc. These same third parties do not mention that shippers can become dependent on individual carriers for high-quality, low-cost service. The results reported here indicate that shippers evaluate the risks and potential costs associated with such dependence. Where these potential costs are too high, shippers will use their own resources rather than depend on outside companies. The continued importance of private fleets shows that carriers have not addressed the risk/dependence issue to many shippers' satisfaction. As shippers reduce the number of carriers used, the importance of perceived risk may very well increase.

This study has tentatively verified that transaction cost factors as well as direct cost and service advantages are involved in the use of private fleets. However, the study population was relatively small and there were only four industries involved. Each shipper reported only relative to its largest customer. Furthermore, important organizational and strategic variables were not considered in the model. Private fleets require knowledgeable managers, who may be in short supply. Company history may also be important to the make-or-buy decision, especially if union employees are involved. Finally, there are many types and sizes of private fleets,(28) and the relative importance of transaction costs probably varies accordingly. Linehaul tractors which cost $100,000 each must be harder to justify based on TCA than local delivery units at $20,000 each.

The relatively small sample size and number of omitted variables suggest two avenues for further research. It would seem appropriate to consider larger scale studies and indepth case analysis. Studies with more respondents could consider all customers served by private fleets to find out whether small customers are treated in the same way as large customers. Also, such studies would allow better investigation of both market effects and industry effects. Indepth case studies could address the importance of specific situational variables and organizational constraints. It would be particularly interesting to determine if functions other than logistics, especially finance, play a major role in the make-or-buy decision for private fleets.

Such studies should lead to a more refined model of private fleet use. Such a model would be invaluable to carriers and other third parties. A quantitative description of the factors leading to private fleet use should allow third-party providers to locate major private fleet operators and design marketing programs to address these users' concerns. It seems likely that private fleets represent the largest single market opportunity available to for-hire motor carriers and logistics service providers. In many cases, continued use of private fleets may be dictated by transaction costs, control issues, etc. In other cases, conversion to third parties may be appropriate. The model envisioned above should help logistics service companies pinpoint the most likely prospects and tailor service offerings to satisfy shipper concerns.


1 1991-1992 Membership Directory, in The Private Carrier 28, no. 6 (June. 1991), 27-63.

2 Robert V. Delaney, "Productivity, Washington, and Money," The Private Carrier 27, no. 11 (November, 1990), 38-41.

3 Same reference as Note 1, pp. 61-63.

4 Lamar Lee, Jr. and Donald W. Dobler, Purchasing and Materials Management, 3d ed., (New York: McGraw-Hill, 1977), pp. 301-315. Gary J. Zenz, Purchasing and the Management of Materials, 6th ed., (New York: John Wiley, 1987), pp. 285-298.

5 For example, George John and Barton A. Weitz, "Forward Integration into Distribution: An Empirical Test of Transaction Cost Analysis," Journal of Law, Organization, and Economics 4, no. 2 (Fall, 1988), 337-355; Gordon Walker and David Weber, "A Transaction Cost Approach to Make-or-Buy Decisions," Administrative Science Quarterly 29, no. 3 (1984), 373-391.

6 David L. Anderson and James Gillies, "Third Party Logistics: What is the Trend?" Annual Conference Proceedings (Council of Logistics Management: Oak Brook, Illinois, 1990), v. 1, 73-82.

7 Bernard J. La Londe, James M. Masters, Arnold B. Maltz, and Lisa R. Williams, The Evolution, Status, and Future of the Corporate Transportation Function (Louisville, Kentucky: American Society of Transportation and Logistics, 1991).

8 Delaney, 40.

9 Michael A. McGinnis, "The Relative Importance of Cost and Service in Freight Transportation Choice: Before and After Deregulation," Transportation Journal 30, no. 1 (Fall, 1990): 12-19.

10 J. Richard Jones, Industrial Shipper Survey Plant Level (Washington, D.C.: U.S. Department of Transportation, 1975).

11 Joseph L. Cavinato, "The Logistics of Contract Manufacturing," International Journal of Physical Distribution and Materials Management 19, no. 1 (1989), 13-20.

12 John Fernie, "Contract Distribution in Multiple Retailing," International Journal of Physical Distribution and Materials Management 19, no. 2 (1989), 14-36.

13 Richard L. Coase, "The Nature of the Firm," Economica 4 (November 1937), 386-405.

14 Oliver E. Williamson, The Economic Institutions of Capitalism (New York: The Free Press, 1985).

15 Frederick J. Beier, "Transportation Contracts and the Experience Effect, A Framework for Future Research," Journal of Business Logistics, Vol. 10, No. 2 (1989), 73-89.

16 Williamson, 1985, p. 95.

17 F. M. Scherer and David Ross, Industrial Market Structure and Economic Performance (Boston: Houghton Mifflin, 1990).

18 Lee and Dobler, 1977, pp. 303-304.

19 Zenz, 1987, pp. 287-293.

20 Thomas W. Speh and James A. Blomquist, The Financial Evaluation of Warehousing Options: An Examination and Appraisal of Contemporary Practices (Oxford, Ohio: Warehouse Research Center, 1988), pp. 17-18.

21 Industry effects on the warehousing make-or-buy decision are evaluated in Arnold B. Maltz, Outsourcing the Corporate Logistics Function: Economic and Strategic Considerations (Ohio State University: unpublished Ph.D. dissertation, 1992).

22 Erin Anderson and Barton A. Weitz, "Make-or-Buy Decisions: Vertical Integration and Marketing Productivity," Sloan Management Review 27, no. 3 (Spring 1986), 3-20.

23 J. Scott Armstrong and Terry S. Overton, "Estimating Non-Response Bias in Mail Surveys," Journal of Marketing Research 14, No. 4 (August, 1977), 396-402.

24 Beier, p. 79.

25 Williamson, p. 96.

26 Ibid., p. 79.

27 Ibid., p. 80.

28 Same reference as Note 1.

The author would like to thank Kevin Boberg, Bernard J. La Londe, and the anonymous reviewers for their suggestions.

Mr. Maltz is assistant professor of marketing and transportation, New Mexico State University, Las Cruces, New Mexico 88003-0001.
COPYRIGHT 1993 American Society of Transportation and Logistics, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1993 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Maltz, Arnold
Publication:Transportation Journal
Date:Mar 22, 1993
Previous Article:Competitive issues in intermodal railroad-truck service.
Next Article:Logistical Excellence: It's Not Business as Usual.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters