Three essays in transportation economics
Date
2024-08
Authors
Soborowicz, Levi J
Major Professor
Advisor
Kim, Donghyuk
Lyn, Gary
Orazem, Peter
Rose, William
Winters, John
Committee Member
Journal Title
Journal ISSN
Volume Title
Publisher
Altmetrics
Abstract
This dissertation is divided into three chapters, each regarding interstate trucking in some manner. The first chapter is an introductory chapter. The second chapter focuses on the truck driver shortage and different policies targeted at increasing the number of available drivers. The final chapter is a model of transportation prices, where carriers and shippers are mismatched geographically. This causes deadheading when a driver moves to another location without freight. The model is augmented with congestion, causing higher costs on more congested roadways.
Chapter One is a descriptive analysis of the interstate and intrastate carrier markets. I use the Census of Motorcarriers from the Motor Carrier Management System (MCMIS) to describe firm entry and exit dynamics, size, and competition. In this chapter, I classify firms into three categories based on information on their activity and legal status. I use this information to further classify firms’ movement between the three categories. This allows me to describe firm dynamics, including exit and entry, as well as firms changing their legal status between interstate or intrastate carriers. I find that the number of interstate carriers has been rising since the Great Recession. In particular, except for 2008 and 2009, the number of carriers has risen each year. Simultaneously, the number of active intrastate carriers has also increased over the sample period. This demonstrates trucking is an expanding market. I estimate entry and exit rates at the county level. The goal of the exercise is to test the relationship between recessions and entry and exit, as well as the importance of other firms with respect to the creation of new firms. I regress the exit or entry rate on the number of incumbents, the size of the labor force, and the unemployment rate by county. I find in general, entry declines with unemployment but increases the larger the county labor force and number of incumbents. With respect to exit, high unemployment rates correlate with more exits. Secondly, counties with more incumbents and a larger labor force demonstrate higher rates of exit. This suggests that the largest counties and counties with the
most incumbents have a high degree of churn in the form of firms entering and exiting. I then focus the analysis on interstate carriers. These are carriers with the authority to operate across state lines. The interstate trucking market is characterized by a mixture of a few extremely large firms and many extremely small firms. The median trucking firm has one truck, one driver, and one trailer. However, the story is very different at the top of the firm size distribution. Among the top .1% of firms, the average firm has 35 tractors, 27 trailers, and 73 drivers. The difference is even more stark at the very top of the distribution, where firms like UPS, FedEx, Swift, Schneider National, and Old Dominion are extremely large firms. I compute the Herfindahl-Hirschman Index (HHI) over time using the MCMIS data using the miles driven or number of tractors as the measure of a firm’s market share. Interstate trucking exhibits a declining HHI over time. The largest decline occurs during the Great Recession. Finally, I conclude with a brief discussion of a truck driver shortage. A shortage of drivers is a major concern in the transportation industry. This miasma of concern over a shortage of truck drivers is sometimes rooted in the fact that drivers are aging, and demand for transportation services is increasing, exemplified by the fact that the total number of miles driven in the US has increased 13% over the sample period. An alternative explanation for a shortage suggests that the largest firms have trouble retaining workers, but the number of drivers in the market is enough to meet demand. In essence, workers are going into business for themselves. In the case of trucking, this is an attractive explanation because the total number of drivers has increased by 36%, so either demand has grown by more than that over the sample period or large firms are unable to retain workers. Using the MCMIS data, I regress the entry of new firms on a number of CDL-qualified drivers at large firms. I find no evidence that the entry of new firms increases as the number of CDL-qualified drivers at large firms within a county increases.
In Chapter Two, I analyze policies to increase the number of truck drivers. I consider three generic policies: increasing compensation, reducing training costs, and allowing 18-20-year-olds to participate in the market. I also consider a fourth counterfactual where I simulate the impact of an aging workforce on the number of truck drivers. To do this, I use multiple datasets: the Current Population Survey, O*Net, IPEDS, and the Occupational Employment and Wage Statistics dataset. I model the occupational choice of workers using a mixed logit model, where workers condition their choice of occupation on the job characteristics of their previous occupation, the observed wages of the occupation they consider, and individual demographic characteristics. Estimates show that workers respond to wage increases in occupations by switching to the occupation with a wage increase at a higher rate. Secondly, I find that older workers and women are less likely to switch occupations. This is critical for estimating the impact of an aging workforce. Using the model estimates, I find that a wage increase of $1,000 increased the number of drivers nationally by about 21,300 drivers. An alternative policy offers a bonus to new drivers, and that policy increases the number of drivers by about 15,500. Differencing the two estimates shows that the wage increase increases retention by about 5,800 workers in the case of an across-the-board wage increase. Reducing training costs has the smallest impact on the number of workers. A $1,000 scholarship for truck driving training programs increases the number of drivers nationally by about 7,000. Finally, the policy with the largest impact is allowing 18-20-year-olds to participate in the interstate trucking market. This counterfactual shows the number of drivers increased by about 47,000 drivers when allowing the young workers to become truck drivers. Finally, an aging workforce is not a major concern for trucking in the short run. By aging the workforce by 5 years the number of truck drivers increase by about 18,800 drivers. This somewhat surprising result is explained by the fact that older workers are less likely to move occupations. Because older workers are less likely to switch occupations, the occupations with the lowest entry rates and highest retention rates gain workers as the workforce ages because incumbent workers are less willing to move occupations. This should alleviate some concerns about a driver shortage being exacerbated by an aging workforce. However, eventually, if no new workers enter the retirement will dominant the impact of the older workers’ increased disutility of switching occupations. Finally, I decompose the counterfactuals into winners and losers. In the case of wage increases the occupations that lose workers are the occupations that are closest to truck driving. These are occupations such as couriers, bus or taxi drivers, rail workers, etc. The final chapter is a model of freight prices and carriers’ location as a function of congestion and deadheading. The location of carriers is a crucial component of the interstate carrier market. In this chapter I analyze the effect of congestion on the amount and location that trucks deadhead(reposition) to. The model is designed to capture the relationship between prices, congestion, and truck locations. To model this environment, I use a dynamic discrete choice model where carriers choose which location to deadhead to. I model the matching process through which carriers are matched with shippers and the prices for shipping freight that are negotiated in these matches. Finally, carriers are faced with endogenous travel costs to model congestion; costs rise as the number of carriers on a route increases. Explicitly, the model works by first matching carriers and shippers who are the demanders for freight services. These matches result in carriers and shippers bargaining over surplus value using a Nash Bargaining Process Nash (1950). Unmatched carriers either decide to deadhead or remain in the current location. Carriers make the decision based on the relative values of each location discounted into the future. I simulate the model and compute the steady-state using three examples. In each example, I find congestion reduces deadheading, but the effect on prices depends on the characteristics of the location carriers can choose between. In general, carriers are faced with a tradeoff, locations that are most profitable have the highest amount of backhauling and so will be more expensive to travel to due to congestion. This leads to carriers deadheading less, which can further exacerbate the mismatch between carriers and shippers. This is one possible explanation for geographic shortages in drivers. I conclude with a discussion of future research in order to estimate the impact of increased congestion on the amount of deadheading and the implications for the mismatch between carriers and shippers.
Series Number
Journal Issue
Is Version Of
Versions
Series
Academic or Administrative Unit
Type
dissertation