
This article was originally published in the Fall 2012 edition of OnAnalytics, published by the Institute for Business Analytics at Indiana University’s Kelley School of Business.
This article focuses on insights from Alfonso Pedraza Martinez an associate professor in Kelley’s department of Operations and Decision Technologies. Find more information about this research in the article, “Vehicle Replacement in the International Committee of the Red Cross“
Transportation represents an enormous overhead cost for humanitarian agencies, second only to personnel. The International Committee of the Red Cross (ICRC) relies on its fleet of 4 x 4 vehicles to transport people and aid across territories that are often in conflict and roads that are rarely paved. Because the reliability of the fleet is critical to the ICRC’s mission, the organization’s vehicle replacement policy is a topic of major importance. With this study, the researchers examine the ICRC’s Standard Replacement Policy (SRP) in the context of the organization’s data on vehicle usage and replacement. They find that the current SRP is not being followed due to a misalignment of incentives between the organization’s headquarters (HQ) in charge of purchasing the vehicles and the national delegations (ND) that use them. Based on the usage and sales data, the researchers propose a new policy yielding cost savings.
Statement of the problem
The ICRC operates a fleet of 1700 4 x 4 vehicles used in more than 80 countries. The organization upholds the manufacturer’s suggested replacement policy of 5 years or 150,000 km, whichever comes first. These specifications, however, refer to commercial use under “normal” conditions, whereas ICRC vehicles are likely to be driven in rural areas of developing countries where roads are rough or nonexistent. The researchers set out to determine 1) whether the National Delegations were following the present SRP and 2) whether the SRP was optimal from a cost perspective given the actual conditions under which the vehicles were being used.
Data Sources Used
The researchers had access to data on ICRC fleets in Afghanistan, Ethiopia, Georgia, and Sudan. The data covered the period from 2002 to 2006 and pertained to procurement and sales, o
perating costs, monthly mileage, and accidents. The researchers also collected qualitative data on vehicle use and replacement practices through interviews with HQ staff and regional- and national-level logisticians.
Analytic Techniques
The nature of the quantitative data collection, which was conducted by ICRC staff in the field under emergency conditions, required a lengthy cleaning process during which the researchers verified each record. The next step was to create descriptive statistics for each of the four categories of quantitative data. The researchers plotted monthly costs as a function of age and monthly mileage over the course of the vehicles’ lifecycles.
To determine whether ND were following the SRP, the researchers plotted the age and odometer of each vehicle at the time of replacement. The drivers of vehicle replacement were determined through a binary logistic model considering the independent variables of age, odometer, and accidents.
For their parameter estimation, the researchers created functions for preventive maintenance and miscellaneous costs and determined the drivers of salvage value using OLS regressions in Stata software. These functions were used to create an optimal replacement model based on a dynamic programming algorithm. The researchers used a C++ application to solve the model. Stochastic simulations were used to check the results’ robustness. Throughout the study, the qualitative information gained from the interviews was used to inform the model.
Results
The descriptive statistics revealed that the vehicles were considerably older than the SRP would prescribe, with median ages of 5 years in three out of four countries. Monthly mileage records showed decreased use over the lifespan of the vehicles despite rigorous standards ensuring continued reliability, which suggested excessive fleet size.
Records of the age and odometer of the vehicles at replacement clearly reveal that ND were not following the SRP. Only 5% of vehicles were sold according to the policy, with more than 50% of vehicles surpassing both 5 years and 150,000 km before being sold.
Combining the statistical analysis with information from the qualitative interviews, the researchers observed that the unique economic drivers of the humanitarian setting were prompting ND to keep the vehicles longer than the SRP prescribed. HQ was responsible for purchasing the vehicles and received the vehicles’ salvage value at the time of sale. ND were responsible for paying depreciation during their usage of the vehicles, but only for the first five years. At that point, the policy indicated that ND should sell the vehicles, but because they were no longer paying a monthly fee to HQ and would not receive the vehicles’ salvage value, there was little incentive to do so. The evidence of fleet inflation further suggested that the cost analysis of HQ logisticians did no
t align with the practical considerations of ND.
Based on the binary logistic model, age and odometer appeared to be the drivers of the replacement decision. The 150,000 km cutoff was the approximate average for odometer, which was the replacement mileage suggested by the manufacturer for commercial fleets, but the corollary age at which ND were replacing vehicles was eight years.
The optimization model, which captured the increase in maintenance costs under the actual conditions in which the vehicles were being driven, the low purchasing cost at which the ICRC was able to procure the vehicles under a special agreement with the manufacturer, and the regression’s result that odometer was the only significant driver of salvage value, revealed that by replacing the vehicles at 100,000 km instead of 150,000 km, the ICRC would save 8.7% on operating costs.
The results were robust to significant variation in purchase price, salvage value, and maintenance costs. Very large increases in maintenance costs or salvage value or very large decreases in purchase cost would indicate earlier replacement. Conversely, very large decreases in maintenance costs or salvage value or very large increases in purchase cost would argue for later replacement.
Although the 100,000 km policy would be considerably more cost effective for the ICRC as a whole, the misalignment of incentives remains between HQ and ND. The researchers note that until the incentives problem is resolved, the new policy may not be implemented at the ND level.
Business Implications
With the ICRC serving as the benchmark in fleet management for other humanitarian organizations such as the World Food Programme, World Vision International, and the International Federation of Red Cross and Red Crescent Societies, this study may serve to influence decision making for a significant portion of the 80,000 4 x 4 vehicles presently in service in the international humanitarian sector.
Moreover, this study demonstrates that policies developed within the commercial sector do not always translate successfully to humanitarian operations. The importance of collecting and analyzing data specific to the field is revealed by the use of the optimization model. This research also demonstrates how quantitative and qualitative data can be combined to offer meaningful insights into policies and their implementation.