Imbalances in the truck market during COVID-19 and their effect on load bundles
The COVID-19 pandemic has shifted freight demand mix and upended carrier networks. In this article, I quantify imbalances in the truck market and their impact on carrier utilization. My analysis shows that Uber Freight load bundles absorbed market imbalances and maintained high carrier utilization.
The impact of the pandemic on the freight market
The COVID-19 pandemic has impacted U.S. freight on both the demand and supply sides. As consumers rushed to buy grocery and hygiene staples, the trucking industry saw a spike in demand to support restocking of essential consumer goods in large metropolitan areas, leading to increases in spot load prices. At the same time, manufacturing and non-essential retail freight volumes have decreased due to shelter-in-place orders, leading to a shift in the typical freight demand mix with overall demand being up in the last month.
Meanwhile, the supply of trucking capacity has experienced conflicting impacts in recent weeks. Some factors have served to increase capacity on the carrier side: shelter-in-place orders have significantly reduced congestion, allowing trucks to travel faster through urban areas. In addition, the Federal Motor Carrier Safety Authority and individual states have relaxed some regulations, including temporarily waiving hours-of-service rules for emergency relief loads, boosting supply by allowing trucks to be on the road longer each day.
On the other hand, some measures to contain the pandemic have had negative effects on carrier capacity. For example, Pennsylvania temporarily closed all highway rest areas, making it harder for truckers to move goods across the state and placing a burden on a group already vulnerable to contracting the virus. As a result, some large carriers pulled drivers out of the market, especially for long haul routes. The increase in demand for shipments together with social distancing measures implemented by shippers at facilities, while understandable, has led to record wait times at trucking facilities.
In addition, the freight demand mix shift might have negatively impacted carriers’ utilization, the percent of time their trucks are full. Utilization is a critical efficiency metric that is likely to be affected by the balance of loads being shipped to and from each regional market. In a perfectly balanced market, there would be an outbound load available for every incoming load; e.g. every truck hauling food to a metropolitan area would return with some manufactured goods, keeping trucks running full both ways. However, spiking demand for certain goods and declining demand for others might leave more carriers without a matching return load. The demand mix shift following the COVID-19 pandemic resulted in increased imbalances between inbound and outbound loads in certain regions. This imbalance could make efficient routing more difficult, thus decreasing utilization and negatively impacting carrier unit economics.
In this blog post, I first demonstrate that the COVID-19 pandemic has resulted in increased imbalance in freight networks in the U.S. I then analyze Uber Freight’s bundles product that is aimed at offering high utilization routes to carriers and show that it has continued to perform well, in spite of the imbalance.
Increased imbalances in freight networks
To measure the imbalance in the freight marketplace, I look at the imbalance or “tightness” of each regional freight market, utilizing DAT’s market definitions. I measure this by the absolute difference between the number of available outbound loads and the number of available inbound loads in each market. “Tight” markets have more outbound loads than inbound loads, while “loose” markets have more inbound loads than outbound loads. The figure below shows how these imbalances played out in recent weeks for a few example markets that experienced sharp changes during the crisis. While the San Diego and New York City areas saw surges in inbound loads without outbound loads to offset them, already tight markets like California’s Inland Empire and the Toledo, OH region experienced further increases in outbound loads.
To judge imbalance in the overall freight sector based on Uber Freight data, I calculate the inbound-outbound delta for each market and then add up the absolute value of these figures over all regional markets for each day. The resulting figure is the number of excess loads on the platform, and an indicator of total market imbalance. In a perfectly balanced freight network, this metric will be zero. The figure below shows that this measure has increased by 68% since the declaration of a national state of emergency on March 13, 2020.
Impact on carrier utilization
Utilization is critical for the success of carriers. Since I do not observe carrier utilization directly from the data, I focus on the bundles feature on the Uber Freight platform. The bundles product algorithmically combines loads to create high utilization routes. In the following, I focus on key statistics on offered bundles to check whether the bundle product is still effective under increased market imbalance.
The overall effect of an imbalanced surge in demand on the bundling feature is not straightforward. On the one hand, the load surge presents more bundling opportunities; for example, there is a higher chance that appointment times of two loads will match and form an efficient bundle. On the other hand, the imbalance means that there are fewer loads per driver available needing to be transported to other markets, thus reducing the probability of returning full. In the following analysis, I look at performance indicators of the Uber Freight bundling feature and how they relate to market imbalances.
First, I look at the share of overall loads offered in bundles to measure the ability of the algorithm to provide efficient route planning. Figure 2 shows a rather flat trend in the weekly percentage of loads bundled. While the share of loads offered as bundles is down by around 6% since the state of emergency declaration, this drop is well within one half of the week-to-week standard deviation.
Second, I consider the route efficiency of the offered bundles as a measure of their quality. Figure 3 plots the average distance between drop-off and pick-up destination for bundles, i.e. the deadhead a carrier would have to incur to reach the next load.¹ The graph remains rather flat, and deadhead even appears to decline starting in early March, falling 4.8% after the state of emergency declaration.
Conclusion
In conclusion, the pandemic-related demand shift upended the freight network, leading to large increases in inbound-outbound imbalances between certain regions. This demand mix shift, however, did not result in large negative effects on the bundling feature. Key performance indices of offered bundles remain within typical volatility before the crisis. This suggests that the bundling feature was able to adjust to the market imbalance caused by the pandemic in the last few weeks, and that algorithmic load matching on the Uber Freight network continued to help carriers maintain high utilization even during times of duress.
Footnotes
[1]:In an earlier publication, I explain in detail how I measure deadhead from the Uber Freight data.
